Build auxiliary in pretty tests
I've encountered a problem with building auxiliary in pretty tests with the following configuration:
```
//@ edition:2021
//@ aux-crate:to_reuse_functions=to-reuse-functions.rs
//@ pretty-mode:hir
//@ pretty-compare-only
//@ pp-exact:delegation-inherit-attributes.pp
```
In such kind of tests, I want to test the resulting HIR, without final typecheck (`pretty-compare-only`). I plan to use such tests during testing the attribute inheritance in delegation feature. However, related auxiliary is built only in [self.typecheck_source](https://github.com/rust-lang/rust/blob/2e667b0c6491678642a83e3aff86626397360af5/src/tools/compiletest/src/runtest/pretty.rs#L79) method (in `compose_and_run_compiler` routine), so it is impossible to execute test in my scenario. As a solution I've added the auxiliary build in [print_source](https://github.com/rust-lang/rust/blob/2e667b0c6491678642a83e3aff86626397360af5/src/tools/compiletest/src/runtest.rs#L434) method, so it is also built in a compare only scenarios.
I've tried to search other solutions to this problem, but I haven't found a way to force auxiliary build in such a configuration of a pretty test.
r? `@petrochenkov`
Make `--print=backend-has-zstd` work by default on any backend
Using a defaulted `CodegenBackend` method that querying for zstd support should automatically print a safe value of `false` on any backend that doesn't specifically indicate the presence or absence of zstd.
This should fix the compiletest failures reported in https://github.com/rust-lang/rust/pull/149666#discussion_r2597881482, which can occur when LLVM is not the default codegen backend.
Correctly provide suggestions when encountering `async fn` with a `dyn Trait` return type
Point at return type of `async fn`s instead of their def `Span`.
Correctly provide suggestions when encountering `async fn` with a `dyn Trait` return type.
Correctly address `type Alias = dyn Trait` when `fn foo() -> Alias { .. }`.
Fixrust-lang/rust#147894.
Remove -Zoom=panic
There are major questions remaining about the reentrancy that this allows. It doesn't have any users on github outside of a single project that uses it in a panic=abort project to show backtraces. It can still be emulated through `#[alloc_error_handler]` or `set_alloc_error_hook` depending on if you use the standard library or not. And finally it makes it harder to do various improvements to the allocator shim.
With this PR the sole remaining symbol in the allocator shim that is not effectively emulating weak symbols is the symbol that prevents skipping the allocator shim on stable even when it would otherwise be empty because libstd + `#[global_allocator]` is used.
Closes https://github.com/rust-lang/rust/issues/43596
Fixes https://github.com/rust-lang/rust/issues/126683
```
error[E0277]: the size for values of type `(dyn Debug + 'static)` cannot be known at compilation time
--> $DIR/dyn-trait-type-alias-return-type.rs:4:11
|
LL | fn f() -> T { loop {} }
| ^ doesn't have a size known at compile-time
|
= help: the trait `Sized` is not implemented for `(dyn Debug + 'static)`
note: this type alias is unsized
--> $DIR/dyn-trait-type-alias-return-type.rs:1:1
|
LL | type T = dyn core::fmt::Debug;
| ^^^^^^
= note: the return type of a function must have a statically known size
```
bootstrap: Use a `CompiletestMode` enum instead of bare strings
This gives better static checking when dealing with compiletest modes in bootstrap, especially for those modes that overlap with the name of a test suite directory.
r? jieyouxu (or bootstrap)
Retire `opt_str2` from compiletest cli parsing
We have `Option<..>`, we don't need to invent "(none)" as option-at-home.
- More specifically, when some test suite expect certain values to be present, they should `.expect(..)` the value, and not potentially receive some "(none)" in some cases.
r? Zalathar
Move the libgccjit.so file in a target directory
Since GCC is not multi-target, we need multiple libgccjit.so. Our solution to have a directory per target so that we can have multiple libgccjit.so.
r? `@Kobzol`
Using a defaulted `CodegenBackend` method that querying for zstd support should
automatically print a safe value of `false` on any backend that doesn't
specifically indicate the presence or absence of zstd.
Subtree sync for rustc_codegen_cranelift
The main highlights this time are a Cranelift update and adding actual support for `-Cdebuginfo=line-tables-only` rather than treating it the same as `-Cdebuginfo=full`.
r? `@ghost`
`@rustbot` label +A-codegen +A-cranelift +T-compiler
This derive was an artifact of test-only method `Cache::all` wanting to
automatically sort its output to hide HashMap iteration order.
We can achieve an equivalent result by requiring the caller to provide a
projection function that returns results that _are_ sortable.
Move ambient gdb discovery from compiletest to bootstrap
- Follow-up to https://github.com/rust-lang/rust/pull/148099
---
This code takes the compiletest code for discovering an “ambient” `gdb` in the user's path, and moves it to bootstrap.
One of the eventual goals is to allow compiletest to assume that if it has been asked to run the debuginfo-gdb suite, then it *must* have been passed an explicit `--gdb`, though we aren't quite there yet.
r? jieyouxu
Tidying up tests/ui/issues tests [3/N]
> [!NOTE]
> Intermediate commits are intended to help review, but will be squashed add comment commit prior to merge.
part of rust-lang/rust#133895
r? Kivooeo
Fix off-by-one staging output when testing the library
It seems generally useful to store the `Mode` in `Cargo`, I remember thinking a few times that it would be useful in other places.
Fixes: https://github.com/rust-lang/rust/issues/149558
r? `@jieyouxu`
Add warn-by-default lint for visibility on `const _` declarations
Add a warn-by-default `unused_visibilities` lint for visibility qualifiers on `const _` declarations—e.g. `pub const _: () = ();`. Such qualifiers have no effect.
A [Sourcegraph search](https://sourcegraph.com/search?q=context:global+lang:Rust+pub%5Cs*%28%5C%28.*%5C%29%29%3F%5Cs*const%5Cs%2B_%5Cs*:&patternType=regexp&case=yes&sm=0) suggests that this pattern is relatively rare, and mostly found in tests (with only 3 exceptions). So perhaps this could become an FCW/hard error in the future.
`@rustbot` label T-lang A-lints A-visibility -T-clippy
While some of this information is needed by debugger discovery, it is also
needed by non-debuginfo tests, so the code doesn't belong in the `debuggers`
module.
Matthias Krüger