Document subdirectories of UI tests with README files
Part of rust-lang/rust#133895 and the [2025 Google Summer of Code](https://blog.rust-lang.org/2025/05/08/gsoc-2025-selected-projects/) associated project.
When adding a new UI test, one is faced with hundreds of subdirectories in `tests/ui` reflecting various categories. Knowing where to put the new test is not trivial, as many of the categories have slightly misleading names. For example, `moves` does not only refer to the `move` keyword but to functions taking ownership in general, whereas `allocator` does not refer to allocation in general but rather to the very specific `allocator_api` and `global_allocator` features.
Many contributors will therefore place their test at the top level of ̀`tests/ui` where it will be mixed with hundreds of unrelated tests.
This PR is a tentative move towards more clearly defined tag/categories, with a SUMMARY.md file documenting the true purpose of each subdirectory, placed inside `tests/ui`.
r? ``@jieyouxu``
Let String pass #[track_caller] to its Vec calls
I've added `#[track_caller]` to `String` methods that delegate to `Vec` methods that already have `#[track_caller]`.
I've also added `#[track_caller]` to methods that have `assert!` or `panic!` due to invalid inputs.
Skip no-op drop glue
Since rust-lang/rust#122662 this no longer gets used in vtables, so we're safe to fully
drop generating functions from vtables. Those are eventually cleaned up
by LLVM, but it's wasteful to produce them in the first place.
This doesn't appear to be a significant win (and shows some slight regressions) but
seems like the right thing to do. At minimum it reduces noise in the LLVM IR we generate,
which seems like a good thing.
Rollup of 10 pull requests
Successful merges:
- rust-lang/rust#142458 (Merge unboxed trait object error suggestion into regular dyn incompat error)
- rust-lang/rust#142593 (Add a warning to LateContext::get_def_path)
- rust-lang/rust#142594 (Add DesugaringKind::FormatLiteral)
- rust-lang/rust#142740 (Clean-up `FnCtxt::is_destruct_assignment_desugaring`)
- rust-lang/rust#142780 (Port `#[must_use]` to new attribute parsing infrastructure)
- rust-lang/rust#142798 (Don't fail to parse a struct if a semicolon is used to separate fields)
- rust-lang/rust#142856 (Add a few inline directives in rustc_serialize.)
- rust-lang/rust#142868 (remove few allow(dead_code))
- rust-lang/rust#142874 (cranelift: fix target feature name typo: "fxsr")
- rust-lang/rust#142877 (Document why tidy checks if `eslint` is installed via `npm`)
r? `@ghost`
`@rustbot` modify labels: rollup
Add a few inline directives in rustc_serialize.
I see `debug_strict_add` and `debug_strict_sub` appearing in callgrind output. This bothers me. This PR should make them disappear.
Clean-up `FnCtxt::is_destruct_assignment_desugaring`
I noticed the docstring thing and fixed that; then I thought I'd rewrite the method using let-chains while I'm at it, since that seemed appropriate. But I don't feel too strongly about the second change, so let me know if I should remove it
Add DesugaringKind::FormatLiteral
Implements `DesugaringKind::FormatLiteral` to mark the FormatArgs desugaring of format literals. The main use for this is to stop yapping about about formatting parameters if we're not anywhere near a format literal. The other use case is to fix suggestions such as https://github.com/rust-lang/rust/issues/141350. It might also be useful for new or existing diagnostics that check whether they're in a format-like macro.
cc `@xizheyin` `@fmease`
Merge unboxed trait object error suggestion into regular dyn incompat error
Another hir-walker removed from the well-formed queries. This error was always a duplicate of another, but it was able to provide more information because it could invoke `is_dyn_compatible` without worrying about cycle errors. That's also the reason we can't put the error directly into hir_ty_lowering when lowering a `dyn Trait` within an associated item signature. So instead I packed it into the error handling of wf obligation checking.
completely deduplicate `Visitor` and `MutVisitor`
r? oli-obk
This closesrust-lang/rust#127615.
### Discussion
> * Give every `MutVisitor::visit_*` method a corresponding `flat_map_*` method.
Not every AST node exists in a location where they can be mapped to multiple instances of themselves. Not every AST node exists in a location where they can be removed from existence (e.g. `filter_map_expr`). I don't think this is doable.
> * Give every `MutVisitor::visit_*` method a corresponding `Visitor` method and vice versa
The only three remaining method-level asymmetries after this PR are `visit_stmt` and `visit_nested_use_tree` (only on `Visitor`) and `visit_span` (only on `MutVisitor`).
`visit_stmt` doesn't seem applicable to `MutVisitor` because `walk_flat_map_stmt_kind` will ask `flat_map_item` / `filter_map_expr` to potentially turn a single `Stmt` to multiple based on what a visitor wants. So only using `flat_map_stmt` seems appropriate.
`visit_nested_use_tree` is used for `rustc_resolve` to track stuff. Not useful for `MutVisitor` for now.
`visit_span` is currently not used for `MutVisitor` already, it was just kept in case we want to revive rust-lang/rust#127241. cc `@cjgillot` maybe we could remove for now and re-insert later if we find a use-case? It does involve some extra effort to maintain.
* Remaining FIXMEs
`visit_lifetime` has an extra param for `Visitor` that's not in `MutVisitor`. This is again something only used by `rustc_resolve`. I think we can keep that symmetry for now.
Enable textrel-on-minimal-lib for Windows
`bin_name` needs to be used when building a runnable executable.
Addresses item in rust-lang/rust#128602
---
try-job: x86_64-mingw-*
try-job: x86_64-msvc-*
try-job: i686-msvc-*
Enable fmt-write-bloat for Windows
Seems to be working fine for MSVC once it has the correct binary name.
Addresses item in rust-lang/rust#128602
---
try-job: x86_64-mingw-*
try-job: x86_64-msvc-*
try-job: i686-msvc-*
All HIR attributes are outer
Fixes https://github.com/rust-lang/rust/issues/142649. Closes https://github.com/rust-lang/rust/pull/142759.
All HIR attributes, including parsed and not yet parsed, will now be rendered as outer attributes by `rustc_hir_pretty`. The original style of the corresponding AST attribute(s) is not relevant for pretty printing, only for diagnostics.
r? ````@jdonszelmann````
rustdoc_json: conversion cleanups
A bunch of clean-to-types conversion cleanups I found while working on perf-related stuff in rustdoc_json.
r? ```@aDotInTheVoid```
improve search graph docs, reset `encountered_overflow` between reruns
I think this shouldn't really matter for now. It will be more relevant for my current rework as we otherwise cannot partially reevaluate the root goal in case there has been overflow during the prervious iteration.
r? ````@BoxyUwU````
Port `#[rustc_pub_transparent]` to the new attribute system
Very similar to rust-lang/rust#142498.
This is a part of rust-lang/rust#131229, so
r? ````@jdonszelmann````
---
For reference, the `#[rustc_pub_transparent]` attribute was created by me back in rust-lang/rust#129487.
As mentioned back in https://github.com/rust-lang/rust/pull/129487#discussion_r1730120385, this attribute does not check that it is applied to an ADT, because it checks that `#[repr(transparent)]` is also applied to the same item, which, in turn, should check for ADT.
Pass -Cpanic=abort for the panic_abort crate
The panic_abort crate must be compiled with panic=abort, but cargo doesn't allow setting the panic strategy for a single crate the usual way using `panic="abort"`, but luckily per-package rustflags do allow this. Bootstrap previously handled this in its rustc wrapper, but for example the build systems of cg_clif and cg_gcc don't use the rustc wrapper, so they would either need to add one, patch the standard library or be unable to build a sysroot suitable for both panic=abort and panic=unwind (as is currently the case).
Required for https://github.com/rust-lang/rustc_codegen_cranelift/issues/1567
forward the bootstrap `runner` to `run-make`
The runner was already forwarded to `compiletest`, this just passes it on to `run-make` and uses it in the `run` functions.
The configuration can look like this
```toml
# in bootstrap.toml
[target.s390x-unknown-linux-gnu]
runner = "qemu-s390x -L /usr/s390x-linux-gnu"
```
Any C compilation automatically sets the correct target. Calls to rustc must use `.target(target())`. Then, a command like below will work by cross-compiling to the given target, and using the given runner for that target to execute the binary:
```
./x test tests/run-make/c-link-to-rust-va-list-fn --target s390x-unknown-linux-gnu
```
The runner can also be used for e.g. running with `valgrind`.
This PR also enables its use in the test case that I care about, hopefully that actually does work on the platforms that CI uses. We should probably run some try jobs to be sure?
r? `@jieyouxu`
try-job: test-various
try-job: armhf-gnu
Avoid a few more allocations in `write_shared.rs`
Inspired by rust-lang/rust#141421 , avoids a few `Vec`, `PathBuf` and `String` allocations in `write_shared.rs`. I don't think these will show up on benchmarks, but are still worthwhile IMHO.
Also includes a few small cleanups.
r? nnethercote - if you'd like :)
Since 122662 this no longer gets used in vtables, so we're safe to fully
drop generating these empty functions. Those are eventually cleaned up
by LLVM, but it's wasteful to produce them in the first place.
This also adds a missing test for fn-ptr casts, which do still need to
generate no-op drop glue. It's possible a future optimization could
point all of those at the same drop glue (e.g., for *mut ()) rather than
for each separate type, but that would require extra work for CFI and
isn't particularly easy to do anyway.
- `convert_static` -> `from_clean_static`
- `from_function` -> `from_clean_function`
To match the pre-existing `from_clean_item` and `FromClean::from_clean`.
I left `JsonRenderer::convert_item` unchanged because it's a bit
different.
The `FromClean` trait is used a lot for converting to rustdoc-json
format. But it's not used universally; there are still some ad hoc
functions and methods for converting. This commit fixes this
inconsistency by using `FromClean` more.
The commit also introduces `FromClean` for `Box` and `Option`. This lets
a lot of `as_ref` and `map` calls be removed in favour of simple
`into_json` calls.
We currently have both `FromClean<clean::Constant> for Constant` and
`FromClean<clean::ConstantKind> for Constant` which are basically
identical, but the former is unused.
