Add unstable `--print=supported-crate-types` option
MCP: https://github.com/rust-lang/compiler-team/issues/836
Tracking issue: https://github.com/rust-lang/rust/issues/138640
### Test coverage
Two tests:
1. `tests/ui/print-request/stability.rs` to check that `--print=supported-crate-types` is `-Zunstable-options`-gated
2. `tests/ui/print-request/supported-crate-types.rs` is added as a basic smoke test. Observe that the compiler stdout corresponds to the below *Example output* section (e.g. `proc-macro` is unsupported on `wasm32-unknown-unknown` currently).
### Example output
<details>
<summary>For `x86_64-unknown-linux-gnu`</summary>
Notice the presence of `{c,}dylib` and `proc-macro`:
```
bin
cdylib
dylib
lib
proc-macro
rlib
staticlib
```
</details>
<details>
<summary>For `wasm32-unknown-unknown`</summary>
Notice the absence of `dylib` and `proc-macro`:
```
bin
cdylib
lib
rlib
staticlib
```
</details>
<details>
<summary>For `x86_64-unknown-linux-musl`</summary>
Notice the absence of `{c,}dylib` but presence of `proc-macro`:
```
bin
lib
proc-macro
rlib
staticlib
```
</details>
### Documentation
I added an entry in the unstable book's print request section to document this `supported-crate-types` print request.
### Unresolved questions
- [ ] (Name bikeshedding) is `supported-crate-types` a good name for the print request? I'm inclined to say it's good enough for an unstable print request, but may be worth revisiting at stabilization time.
### Stability
This print request being added is *unstable* in this PR. A separate stabilization PR following the usual compiler flag stabilization procedure should be filed for stabilization after some baking time.
### Review remarks
Best reviewed commit-by-commit.
r? compiler
Move some driver code around
`--emit mir`, `#[rustc_symbol_name]` and `#[rustc_def_path]` now run before codegen and thus work even if codegen fails. This can help with debugging.
Reduce FormattingOptions to 64 bits
This is part of https://github.com/rust-lang/rust/issues/99012
This reduces FormattingOptions from 6-7 machine words (384 bits on 64-bit platforms, 224 bits on 32-bit platforms) to just 64 bits (a single register on 64-bit platforms).
Before:
```rust
pub struct FormattingOptions {
flags: u32, // only 6 bits used
fill: char,
align: Option<Alignment>,
width: Option<usize>,
precision: Option<usize>,
}
```
After:
```rust
pub struct FormattingOptions {
/// Bits:
/// - 0-20: fill character (21 bits, a full `char`)
/// - 21: `+` flag
/// - 22: `-` flag
/// - 23: `#` flag
/// - 24: `0` flag
/// - 25: `x?` flag
/// - 26: `X?` flag
/// - 27: Width flag (if set, the width field below is used)
/// - 28: Precision flag (if set, the precision field below is used)
/// - 29-30: Alignment (0: Left, 1: Right, 2: Center, 3: Unknown)
/// - 31: Always set to 1
flags: u32,
/// Width if width flag above is set. Otherwise, always 0.
width: u16,
/// Precision if precision flag above is set. Otherwise, always 0.
precision: u16,
}
```
Since array and slice constants are now lowered to array and slice
patterns, `non_scalar_compare` was only called for string comparisons.
This specializes it to strings, renames it, and removes the unused
array-unsizing logic.
This also updates some outdated doc comments.
An assignment such as
(a, b) = (b, c);
desugars to the HIR
{ let (lhs, lhs) = (b, c); a = lhs; b = lhs; };
The repeated `lhs` leads to multiple Locals assigned to the same DILocalVariable. Rather than
attempting to fix that, get rid of the debug info for these bindings that don't even exist
in the program to begin with.
Fixes#138198
Avoid no-op unlink+link dances in incr comp
Incremental compilation scales quite poorly with the number of CGUs. This PR improves one reason for that.
The incr comp process hard-links all the files from an old session into a new one, then it runs the backend, which may just hard-link the new session files into the output directory. Then codegen hard-links all the output files back to the new session directory.
This PR (perhaps unimaginatively) fixes the silliness that ensues in the last step. The old `link_or_copy` implementation would be passed pairs of paths which are already the same inode, then it would blindly delete the destination and re-create the hard-link that it just deleted. This PR lets us skip both those operations. We don't skip the other two hard-links.
`cargo +stage1 b && touch crates/core/main.rs && strace -cfw -elink,linkat,unlink,unlinkat cargo +stage1 b` before and then after on `ripgrep-13.0.0`:
```
% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
52.56 0.024950 25 978 485 unlink
34.38 0.016318 22 727 linkat
13.06 0.006200 24 249 unlinkat
------ ----------- ----------- --------- --------- ----------------
100.00 0.047467 24 1954 485 total
```
```
% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
42.83 0.014521 57 252 unlink
38.41 0.013021 26 486 linkat
18.77 0.006362 25 249 unlinkat
------ ----------- ----------- --------- --------- ----------------
100.00 0.033904 34 987 total
```
This reduces the number of hard-links that are causing perf troubles, noted in https://github.com/rust-lang/rust/issues/64291 and https://github.com/rust-lang/rust/issues/137560
Add an attribute that makes the spans from a macro edition 2021, and fix pin on edition 2024 with it
Fixes a regression, see issue below. This is a temporary fix, super let is the real solution.
Closes#138596
add `naked_functions_target_feature` unstable feature
tracking issue: https://github.com/rust-lang/rust/issues/138568
tagging https://github.com/rust-lang/rust/pull/134213https://github.com/rust-lang/rust/issues/90957
This PR puts `#[target_feature(/* ... */)]` on `#[naked]` functions behind its own feature gate, so that naked functions can be stabilized. It turns out that supporting `target_feature` on naked functions is tricky on some targets, so we're splitting it out to not block stabilization of naked functions themselves. See the tracking issue for more information and workarounds.
Note that at the time of writing, the `target_features` attribute is ignored when generating code for naked functions.
r? ``@Amanieu``
They are marked with hidden visibility to prevent them from getting
exported, so we shouldn't ask the linker to export them anyway. The only
thing that does it cause a warning on macOS.
Computing the current dll path is somewhat expensive relative to other
work when compiling `fn main() {}` as `dladdr` needs to iterate over the
symbol table of librustc_driver.so until it finds a match.
Handle spans of `~const`, `const` and `async` trait bounds in macro expansion
r? `@compiler-errors`
`visit_span` is actually only used in one place (the `transcribe::Marker`), and all of this syntax is unstable, so while it would still be nice to write a test for it, I wager there's lots more interesting things in `transcribe::Marker` to write tests for. And the worst is some diagnostics being weird or incremental being not as incremental as it could be
interpret memory access hooks: also pass through the Pointer used for the access
In some ongoing work on the Miri side, we need the absolute address that the memory access occurred at. That is non-trivial to obtain since we don't have an `ecx`. So pass through the `Pointer` used for the access, which contains the address, and which is available everywhere we are calling these hooks.
r? `@oli-obk`
Rollup of 8 pull requests
Successful merges:
- #138435 (Add support for postfix yield expressions)
- #138685 (Use `Option<Ident>` for lowered param names.)
- #138700 (Suggest `-Whelp` when pass `--print lints` to rustc)
- #138727 (Do not rely on `type_var_origin` in `OrphanCheckErr::NonLocalInputType`)
- #138729 (Clean up `FnCtxt::resolve_coroutine_interiors`)
- #138731 (coverage: Add LLVM plumbing for expansion regions)
- #138732 (Use `def_path_str` for def id arg in `UnsupportedOpInfo`)
- #138735 (Remove `llvm` and `llvms` triagebot ping aliases for `icebreakers-llvm` ping group)
r? `@ghost`
`@rustbot` modify labels: rollup
expand: Leave traces when expanding `cfg_attr` attributes
Currently `cfg_trace` just disappears during expansion, but after this PR `#[cfg_attr(some tokens)]` will leave a `#[cfg_attr_trace(some tokens)]` attribute instead of itself in AST after expansion (the new attribute is built-in and inert, its inner tokens are the same as in the original attribute).
This trace attribute can then be used by lints or other diagnostics, #133823 has some examples.
Tokens in these trace attributes are set to an empty token stream, so the traces are non-existent for proc macros and cannot affect any user-observable behavior.
This is also a weakness, because if a proc macro processes some code with the trace attributes, they will be lost, so the traces are best effort rather than precise.
The next step is to do the same thing with `cfg` attributes (`#[cfg(TRUE)]` currently remains in both AST and tokens after expanding, it should be replaced with a trace instead).
The idea belongs to `@estebank.`
Use `def_path_str` for def id arg in `UnsupportedOpInfo`
We could alternatively just omit the def path from the label, but I think it's fine to keep around
Fixes#138730
coverage: Add LLVM plumbing for expansion regions
This is currently unused, but paves the way for future work on expansion regions without having to worry about the FFI parts.
The span conversion refactoring is only loosely related, but I've included it here because it would conflict with the main changes in `fill_region_tables`, and is pretty straightforward on its own.
Do not rely on `type_var_origin` in `OrphanCheckErr::NonLocalInputType`
The ordering of ty var unification means that we may end up with a root variable whose ty var origin is from another item's params.
Let's not rely on this by just unifying the infer vars with the params of the impl + resolving. It's kinda goofy but it's clearer IMO.
Fixes#132826.
r? `@fmease` or `@lcnr`
Use `Option<Ident>` for lowered param names.
Parameter patterns are lowered to an `Ident` by `lower_fn_params_to_names`, which is used when lowering bare function types, trait methods, and foreign functions. Currently, there are two exceptional cases where the lowered param can become an empty `Ident`.
- If the incoming pattern is an empty `Ident`. This occurs if the parameter is anonymous, e.g. in a bare function type.
- If the incoming pattern is neither an ident nor an underscore. Any such parameter will have triggered a compile error (hence the `span_delayed_bug`), but lowering still occurs.
This commit replaces these empty `Ident` results with `None`, which eliminates a number of `kw::Empty` uses, and makes it impossible to fail to check for these exceptional cases.
Note: the `FIXME` comment in `is_unwrap_or_empty_symbol` is removed. It actually should have been removed in #138482, the precursor to this PR. That PR changed the lowering of wild patterns to `_` symbols instead of empty symbols, which made the mentioned underscore check load-bearing.
r? ``@compiler-errors``
Add support for postfix yield expressions
We've been having a discussion about whether we want postfix yield, or want to stick with prefix yield, or have both. I figured it's easy enough to support both for now and let us play around with them while the feature is still experimental.
This PR treats `yield x` and `x.yield` as semantically equivalent. There was a suggestion to make `yield x` have a `()` type (so it only works in coroutines with `Resume = ()`. I think that'd be worth trying, either in a later PR, or before this one merges, depending on people's opinions.
#43122
Consider fields to be inhabited if they are unstable
Fixes#133885 with a simple heuristic
r? Nadrieril
Not totally certain if this needs T-lang approval or a crater run.
For expansion region support, we will want to be able to convert and check
spans before creating a corresponding local file ID.
If we create local file IDs eagerly, but some expansion turns out to have no
successfully-converted spans, LLVM will complain about that expansion's file ID
having no regions.
It's very useful. There are some false positives involving integration
tests in `rustc_pattern_analysis` and `rustc_serialize`. There is also a
false positive involving `rustc_driver_impl`'s
`rustc_randomized_layouts` feature. And I removed a `rustc_span` mention
in a doc comment in `rustc_log` because it wasn't integral to the
comment but caused a dev-dependency.
Rollup of 7 pull requests
Successful merges:
- #135394 (`MaybeUninit` inherent slice methods part 2)
- #137051 (Implement default methods for `io::Empty` and `io::Sink`)
- #138001 (mir_build: consider privacy when checking for irrefutable patterns)
- #138540 (core/slice: Mark some `split_off` variants unstably const)
- #138589 (If a label is placed on the block of a loop instead of the header, suggest moving it to the header.)
- #138594 (Fix next solver handling of shallow trait impl check)
- #138613 (Remove E0773 "A builtin-macro was defined more than once.")
Failed merges:
- #138602 (Slim `rustc_parse_format` dependencies down)
r? `@ghost`
`@rustbot` modify labels: rollup
Remove E0773 "A builtin-macro was defined more than once."
Error E0773 "A builtin-macro was defined more than once" is triggered when using the same `#[rustc_builtin_macro(..)]` twice. However, it can only be triggered in unstable code (using a `rustc_` attribute), and there doesn't seem to be any harm in using the same implementation from `compiler/rustc_builtin_macros/…` for multiple macro definitions.
By changing the Box to an Arc in `SyntaxExtensionKind`, we can throw away the `BuiltinMacroState::{NotYetSeen, AlreadySeen}` logic, simplifying things.
Fix next solver handling of shallow trait impl check
I'm trying to remove unnecessary direct calls to `select`, and this one seemed like a good place to start 😆
r? `@compiler-errors` or `@lcnr`
If a label is placed on the block of a loop instead of the header, suggest moving it to the header.
Fixes#138585
If a label is placed on the block of a loop instead of the header, suggest to the user moving it to the loop header instead of ~~suggesting to remove it~~ emitting a tool-only suggestion to remove it.
```rs
fn main() {
loop 'a: { return; }
}
```
```diff
error: block label not supported here
--> src/main.rs:2:10
|
2 | loop 'a: { return; }
| ^^^ not supported here
+ |
+help: if you meant to label the loop, move this label before the loop
+ |
+2 - loop 'a: { return; }
+2 + 'a: loop { return; }
+ |
```
Questions for reviewer:
* The "desired output" in the linked issue had the main diagnostic be "misplaced loop label". Should the main diagnostic message the changed instead of leaving it as "block label not supported here"?
* Should this be `Applicability::MachineApplicable`?
mir_build: consider privacy when checking for irrefutable patterns
This PR fixes#137999.
Note that, since this makes the compiler reject code that was previously accepted, it will probably need a crater run.
I include a commit that factors out a common code pattern into a helper function, purely because the fact that this was repeated all over the place was bothering me. Let me know if I should split that into a separate PR instead.
Represent diagnostic side effects as dep nodes
This changes diagnostic to be tracked as a special dep node (`SideEffect`) instead of having a list of side effects associated with each dep node. `SideEffect` is always red and when forced, it emits the diagnostic and marks itself green. Each emitted diagnostic generates a new `SideEffect` with an unique dep node index.
Some implications of this:
- Diagnostic may now be emitted more than once as they can be emitted once when the `SideEffect` gets marked green and again if the task it depends on needs to be re-executed due to another node being red. It relies on deduplicating of diagnostics to avoid that.
- Anon tasks which emits diagnostics will no longer *incorrectly* be merged with other anon tasks.
- Reusing a CGU will now emit diagnostics from the task generating it.
Parameter patterns are lowered to an `Ident` by
`lower_fn_params_to_names`, which is used when lowering bare function
types, trait methods, and foreign functions. Currently, there are two
exceptional cases where the lowered param can become an empty `Ident`.
- If the incoming pattern is an empty `Ident`. This occurs if the
parameter is anonymous, e.g. in a bare function type.
- If the incoming pattern is neither an ident nor an underscore. Any
such parameter will have triggered a compile error (hence the
`span_delayed_bug`), but lowering still occurs.
This commit replaces these empty `Ident` results with `None`, which
eliminates a number of `kw::Empty` uses, and makes it impossible to fail
to check for these exceptional cases.
Note: the `FIXME` comment in `is_unwrap_or_empty_symbol` is removed. It
actually should have been removed in #138482, the precursor to this PR.
That PR changed the lowering of wild patterns to `_` symbols instead of
empty symbols, which made the mentioned underscore check load-bearing.
Remove existing AFIDT implementation
This experiment will need to be reworked differently; I don't think we'll be going with the `dyn* Future` approach that is currently implemented.
r? oli-obk
Fixes#136286Fixes#137706Fixes#137895
Tracking:
* #133119
Revert: Add *_value methods to proc_macro lib
This reverts https://github.com/rust-lang/rust/pull/136355. That PR caused unexpected breakage:
- the rustc-dev component can no longer be loaded by cargo, which impacts Miri and clippy and likely others
- rustc_lexer can no longer be published to crates.io, which impacts RA
See https://github.com/rust-lang/rust/issues/138647 for context.
Cc `@GuillaumeGomez` `@Amanieu`
rust-lang/rust@23032f31c9 accidentally introduced some nondeterminism
in the ordering of lib.rmeta files, which we caught in our bazel-based
builds only recently due to being further behind than normal. In my
testing, this fixes the issue.
Add `#[cfg(test)]` for Transition in dfa in `rustc_transmute`
`Transition` is only used in the `Transitions::insert` in test after #137776
Detected by #128637
Extract `for_each_immediate_subpat` from THIR pattern visitors
This is extracted from some larger changes I've been working on, trying to introduce a “THIR pattern id” to refer to THIR pattern nodes without a direct reference.
The future of those changes is somewhat uncertain, due to some [proposed changes involving upvar inference](https://rust-lang.zulipchat.com/#narrow/channel/131828-t-compiler/topic/upvar.20inference.20on.20THIR.3F). So I'm taking my preparatory changes that make sense on their own, and extracting them into one or more independent PRs.
---
This particular patch takes two different functions that were both matching on `PatKind` to traverse subpatterns, and extracts the core match into a single helper function.
Rollup of 7 pull requests
Successful merges:
- #138384 (Move `hir::Item::ident` into `hir::ItemKind`.)
- #138508 (Clarify "owned data" in E0515.md)
- #138531 (Store test diffs in job summaries and improve analysis formatting)
- #138533 (Only use `DIST_TRY_BUILD` for try jobs that were not selected explicitly)
- #138556 (Fix ICE: attempted to remap an already remapped filename)
- #138608 (rustc_target: Add target feature constraints for LoongArch)
- #138619 (Flatten `if`s in `rustc_codegen_ssa`)
r? `@ghost`
`@rustbot` modify labels: rollup
Implement `make_mir_visitor` macro to generate `MirVisitor` and `MutMirVisitor`.
Add `ret_local_mut()`, `arg_locals_mut()` and `inner_locals_mut()` to `Body`, specifically for `MutMirVisitor`.
Mangle rustc_std_internal_symbols functions
This reduces the risk of issues when using a staticlib or rust dylib compiled with a different rustc version in a rust program. Currently this will either (in the case of staticlib) cause a linker error due to duplicate symbol definitions, or (in the case of rust dylibs) cause rustc_std_internal_symbols functions to be silently overridden. As rust gets more commonly used inside the implementation of libraries consumed with a C interface (like Spidermonkey, Ruby YJIT (curently has to do partial linking of all rust code to hide all symbols not part of the C api), the Rusticl OpenCL implementation in mesa) this is becoming much more of an issue. With this PR the only symbols remaining with an unmangled name are rust_eh_personality (LLVM doesn't allow renaming it) and `__rust_no_alloc_shim_is_unstable`.
Helps mitigate https://github.com/rust-lang/rust/issues/104707
try-job: aarch64-gnu-debug
try-job: aarch64-apple
try-job: x86_64-apple-1
try-job: x86_64-mingw-1
try-job: i686-mingw-1
try-job: x86_64-msvc-1
try-job: i686-msvc-1
try-job: test-various
try-job: armhf-gnu
Clarify "owned data" in E0515.md
This clarifies the explanation of why this is not allowed and also what to do instead.
Fixes#62071
PS There was suggestion of adding a link to the book. I did not yet do that, but if desired that could be added.
Move `hir::Item::ident` into `hir::ItemKind`.
`hir::Item` has an `ident` field.
- It's always non-empty for these item kinds: `ExternCrate`, `Static`, `Const`, `Fn`, `Macro`, `Mod`, `TyAlias`, `Enum`, `Struct`, `Union`, Trait`, TraitAalis`.
- It's always empty for these item kinds: `ForeignMod`, `GlobalAsm`, `Impl`.
- For `Use`, it is non-empty for `UseKind::Single` and empty for `UseKind::{Glob,ListStem}`.
All of this is quite non-obvious; the only documentation is a single comment saying "The name might be a dummy name in case of anonymous items". Some sites that handle items check for an empty ident, some don't. This is a very C-like way of doing things, but this is Rust, we have sum types, we can do this properly and never forget to check for the exceptional case and never YOLO possibly empty identifiers (or possibly dummy spans) around and hope that things will work out.
This is step towards `kw::Empty` elimination (#137978).
r? `@fmease`
`hir::Item` has an `ident` field.
- It's always non-empty for these item kinds: `ExternCrate`, `Static`,
`Const`, `Fn`, `Macro`, `Mod`, `TyAlias`, `Enum`, `Struct`, `Union`,
Trait`, TraitAalis`.
- It's always empty for these item kinds: `ForeignMod`, `GlobalAsm`,
`Impl`.
- For `Use`, it is non-empty for `UseKind::Single` and empty for
`UseKind::{Glob,ListStem}`.
All of this is quite non-obvious; the only documentation is a single
comment saying "The name might be a dummy name in case of anonymous
items". Some sites that handle items check for an empty ident, some
don't. This is a very C-like way of doing things, but this is Rust, we
have sum types, we can do this properly and never forget to check for
the exceptional case and never YOLO possibly empty identifiers (or
possibly dummy spans) around and hope that things will work out.
The commit is large but it's mostly obvious plumbing work. Some notable
things.
- A similar transformation makes sense for `ast::Item`, but this is
already a big change. That can be done later.
- Lots of assertions are added to item lowering to ensure that
identifiers are empty/non-empty as expected. These will be removable
when `ast::Item` is done later.
- `ItemKind::Use` doesn't get an `Ident`, but `UseKind::Single` does.
- `lower_use_tree` is significantly simpler. No more confusing `&mut
Ident` to deal with.
- `ItemKind::ident` is a new method, it returns an `Option<Ident>`. It's
used with `unwrap` in a few places; sometimes it's hard to tell
exactly which item kinds might occur. None of these unwraps fail on
the test suite. It's conceivable that some might fail on alternative
input. We can deal with those if/when they happen.
- In `trait_path` the `find_map`/`if let` is replaced with a loop, and
things end up much clearer that way.
- `named_span` no longer checks for an empty name; instead the call site
now checks for a missing identifier if necessary.
- `maybe_inline_local` doesn't need the `glob` argument, it can be
computed in-function from the `renamed` argument.
- `arbitrary_source_item_ordering::check_mod` had a big `if` statement
that was just getting the ident from the item kinds that had one. It
could be mostly replaced by a single call to the new `ItemKind::ident`
method.
- `ItemKind` grows from 56 to 64 bytes, but `Item` stays the same size,
and that's what matters, because `ItemKind` only occurs within `Item`.
change config.toml to bootstrap.toml
Currently, both Bootstrap and Cargo uses same name as their configuration file, which can be confusing. This PR is based on a discussion to rename `config.toml` to `bootstrap.toml` for Bootstrap. Closes: https://github.com/rust-lang/rust/issues/126875.
I have split the PR into atomic commits to make it easier to review. Once the changes are finalized, I will squash them. I am particularly concerned about the changes made to modules that are not part of Bootstrap. How should we handle those changes? Should we ping the respective maintainers?
Emit function declarations for functions with `#[linkage="extern_weak"]`
Currently, when declaring an extern weak function in Rust, we use the following syntax:
```rust
unsafe extern "C" {
#[linkage = "extern_weak"]
static FOO: Option<unsafe extern "C" fn() -> ()>;
}
```
This allows runtime-checking the extern weak symbol through the Option.
When emitting LLVM-IR, the Rust compiler currently emits this static as an i8, and a pointer that is initialized with the value of the global i8 and represents the nullabilty e.g.
```
`@FOO` = extern_weak global i8
`@_rust_extern_with_linkage_FOO` = internal global ptr `@FOO`
```
This approach does not work well with CFI, where we need to attach CFI metadata to a concrete function declaration, which was pointed out in https://github.com/rust-lang/rust/issues/115199.
This change switches to emitting a proper function declaration instead of a global i8. This allows CFI to work for extern_weak functions. Example:
```
`@_rust_extern_with_linkage_FOO` = internal global ptr `@FOO`
...
declare !type !61 !type !62 !type !63 !type !64 extern_weak void `@FOO(double)` unnamed_addr #6
```
We keep initializing the Rust internal symbol with the function declaration, which preserves the correct behavior for runtime checking the Option.
r? `@rcvalle`
cc `@jakos-sec`
try-job: test-various
mir_build: Avoid some useless work when visiting "primary" bindings
While looking over `visit_primary_bindings`, I noticed that it does a bunch of extra work to build up a collection of “user-type projections”, even though 2/3 of its call sites don't even use them. Those callers can get the same result via `thir::Pat::walk_always`.
(And it turns out that doing so also avoids creating some redundant user-type entries in MIR for some binding constructs.)
I also noticed that even when the user-type projections *are* used, the process of building them ends up eagerly cloning some nested vectors at every recursion step, even in cases where they won't be used because the current subpattern has no bindings. To avoid this, the visit method now assembles a linked list on the stack containing the information that *would* be needed to create projections, and only creates the concrete projections as needed when a primary binding is encountered.
Some relevant prior PRs:
- #55274
- 0bfe184b1a in #55937
---
There should be no user-visible change in compiler output.
Denote `ControlFlow` as `#[must_use]`
I've repeatedly hit bugs in the compiler due to `ControlFlow` not being marked `#[must_use]`. There seems to be an accepted ACP to make the type `#[must_use]` (https://github.com/rust-lang/libs-team/issues/444), so this PR implements that part of it.
Most of the usages in the compiler that trigger this new warning are "root" usages (calling into an API that uses control-flow internally, but for which the callee doesn't really care) and have been suppressed by `let _ = ...`, but I did legitimately find one instance of a missing `?` and one for a never-used `ControlFlow` value in #137448.
Presumably this needs an FCP too, so I'm opening this and nominating it for T-libs-api.
This PR also touches the tools (incl. rust-analyzer), but if this went into FCP, I'd split those out into separate PRs which can land before this one does.
r? libs-api
`@rustbot` label: T-libs-api I-libs-api-nominated
Stabilize `asm_goto` feature gate
Stabilize `asm_goto` feature (tracked by #119364). The issue will remain open and be updated to track `asm_goto_with_outputs`.
Reference PR: https://github.com/rust-lang/reference/pull/1693
# Stabilization Report
This feature adds a `label <block>` operand type to `asm!`. `<block>` must be a block expression with type unit or never. The address of the block is substituted and the assembly may jump to the block. When block completes the `asm!` block returns and continues execution.
The block starts a new safety context and unsafe operations within must have additional `unsafe`s; the effect of `unsafe` that surrounds `asm!` block is cancelled. See https://github.com/rust-lang/rust/issues/119364#issuecomment-2316037703 and https://github.com/rust-lang/rust/pull/131544.
It's currently forbidden to use `asm_goto` with output operands; that is still unstable under `asm_goto_with_outputs`.
Example:
```rust
unsafe {
asm!(
"jmp {}",
label {
println!("Jumped from asm!");
}
);
}
```
Tests:
- tests/ui/asm/x86_64/goto.rs
- tests/ui/asm/x86_64/goto-block-safe.stderr
- tests/ui/asm/x86_64/bad-options.rs
- tests/codegen/asm/goto.rs
Improve upvar analysis for deref of child capture
Two fixes to the heuristic I implemented in #123660. As I noted in the code:
> Luckily, if this function is not correct, then the program is not unsound, since we still borrowck and validate the choices made from this function -- the only side-effect is that the user may receive unnecessary borrowck errors.
This indeed fixes unnecessary borrowck errors.
r? oli-obk
---
The heuristic is only valid if we deref a `&T`, not a `&mut T` or `Box<T>`, so make sure to check the type. This fixes:
```rust
struct Foo { precise: i32 }
fn mut_ref_inside_mut(f: &mut Foo) {
let x: impl AsyncFn() = async move || {
let y = &f.precise;
};
}
```
Since the capture from `f` to `&f.precise` needs to be treated as a lending borrow from the parent coroutine-closure to the child coroutine.
---
The heuristic is also valid if *any* deref projection in the child capture's projections is a `&T`, but we were only looking at the last one. This ensures that this function is considered not to be lending:
```rust
struct Foo { precise: i32 }
fn ref_inside_mut(f: &mut &Foo) {
let x: impl Fn() -> _ = async move || {
let y = &f.precise;
};
}
```
(Specifically, checking that `impl Fn() -> _` is satisfied is exercising that the coroutine is not considered to be lending.)
Currently, when declaring an extern weak function in Rust, we use the
following syntax:
```rust
unsafe extern "C" {
#[linkage = "extern_weak"]
static FOO: Option<unsafe extern "C" fn() -> ()>;
}
```
This allows runtime-checking the extern weak symbol through the Option.
When emitting LLVM-IR, the Rust compiler currently emits this static
as an i8, and a pointer that is initialized with the value of the global
i8 and represents the nullabilty e.g.
```
@FOO = extern_weak global i8
@_rust_extern_with_linkage_FOO = internal global ptr @FOO
```
This approach does not work well with CFI, where we need to attach CFI
metadata to a concrete function declaration, which was pointed out in
https://github.com/rust-lang/rust/issues/115199.
This change switches to emitting a proper function declaration instead
of a global i8. This allows CFI to work for extern_weak functions.
We keep initializing the Rust internal symbol with the function
declaration, which preserves the correct behavior for runtime checking
the Option.
Co-authored-by: Jakob Koschel <jakobkoschel@google.com>
`lower_generic_bound_predicate` calls `lower_ident`, and then passes the
lowered ident into `new_named_lifetime`, which lowers it again. This
commit avoids the first lowering. This requires adding a `lower_ident`
call on a path that doesn't involve `new_named_lifetime`.
`LoweringContext::new_named_lifetime` lowers the `ident` passed in. Both
of its call sites *also* lower `ident` *before* passing it in. I.e. both
call sites cause the ident to be lowered twice. This commit removes the
lowering at the two call sites, so the ident is only lowered once.
Misc print request handling cleanups + a centralized test for print request stability gating
I was working on implementing `--print=supported-crate-types`, then I noticed some things that were mildly annoying me, so I pulled out these changes. In this PR:
- First commit adds a centralized test `tests/ui/print/stability.rs` that is responsible for exercising stability gating of the print requests.
- AFAICT we didn't have any test that systematically checks this.
- I coalesced `tests/ui/feature-gates/feature-gate-print-check-cfg.rs` (for `--print=check-cfg`) into this test too, since `--print=check-cfg` is only `-Z unstable-options`-gated like other unstable print requests, and is not additionally feature-gated. cc ``@Urgau`` in case you have any concerns.
- Second commit alphabetically sorts the `PrintKind` enum for consistency because the `PRINT_KINDS` list (using the enum) is *already* alphabetically sorted.
- Third commit pulls out two helpers:
1. A helper `check_print_request_stability` for checking stability of print requests and the diagnostics for using unstable print requests without `-Z unstable-options`, to avoid repeating the same logic over and over.
2. A helper `emit_unknown_print_request_help` for the unknown print request diagnostics to make print request collection control flow more obvious.
- Fourth commit renames `PrintKind::{TargetSpec,AllTargetSpecs}` to `PrintKind::{TargetSpecJson,AllTargetSpecsJson}` to better reflect their actual print names, `--print={target-spec-json,all-target-specs-json}`.
r? ``@nnethercote`` (or compiler/reroll)
To correspond to their actual print request names, `target-spec-json`
and `all-target-specs-json`, and for consistency with other print name
<-> print kind mappings.
Remove `#[cfg(not(test))]` gates in `core`
These gates are unnecessary now that unit tests for `core` are in a separate package, `coretests`, instead of in the same files as the source code. They previously prevented the two `core` versions from conflicting with each other.
Add RTN support to rustdoc
This adds support to rustdoc and rustdoc-json for rendering `(..)` RTN (return type notation) style generics.
---
Cleaning `rustc_middle::ty::Ty` is not correct still, though, and ends up rendering a function like:
```rust
pub fn foreign<T: Foreign<bar(..): Send>>()
where
<T as Foreign>::bar(..): 'static,
T::bar(..): Sync,
```
Into this:
```rust
pub fn foreign<T>()
where
T: Foreign,
impl Future<Output = ()>: Send + 'static + Sync,
```
This is because `clean_middle_ty` doesn't actually have sufficient context about whether the RPITIT is in its "defining scope" or not, so we don't know if the type was originally written like `-> impl Trait` or with RTN like `T::method(..)`.
Partially addresses #123996 (i.e., HIR side, not middle::ty one)
The existing method does some non-obvious extra work to collect user types and
build user-type projections, which is specifically needed by `declare_bindings`
and not by the other two callers.
Rollup of 5 pull requests
Successful merges:
- #138283 (Enforce type of const param correctly in MIR typeck)
- #138439 (feat: check ARG_MAX on Unix platforms)
- #138502 (resolve: Avoid some unstable iteration)
- #138514 (Remove fake borrows of refs that are converted into non-refs in `MakeByMoveBody`)
- #138524 (Mark myself as unavailable for reviews temporarily)
r? `@ghost`
`@rustbot` modify labels: rollup
Remove fake borrows of refs that are converted into non-refs in `MakeByMoveBody`
Remove fake borrows of closure captures if that capture has been replaced with a by-move version of that capture.
For example, given an async closure that looks like:
```
let f: Foo;
let c = async move || {
match f { ... }
};
```
... in this pair of coroutine-closure + coroutine, we capture `Foo` in the parent and `&Foo` in the child. We will emit two fake borrows like:
```
_2 = &fake shallow (*(_1.0: &Foo));
_3 = &fake shallow (_1.0: &Foo);
```
However, since the by-move-body transform is responsible for replacing `_1.0: &Foo` with `_1.0: Foo` (since the `AsyncFnOnce` coroutine will own `Foo` by value), that makes the second fake borrow obsolete since we never have an upvar of type `&Foo`, and we should replace it with a `nop`.
As a side-note, we don't actually even care about fake borrows here at all since they're fully a MIR borrowck artifact, and we don't need to borrowck by-move MIR bodies. But it's best to preserve as much as we can between these two bodies :)
Fixes#138501
r? oli-obk
feat: check ARG_MAX on Unix platforms
On Unix the limits can be gargantuan anyway so we're pretty unlikely to hit them, but might still exceed it.
We consult ARG_MAX here to get an estimate.
Fixes#138421
r? `@jieyouxu`
Enforce type of const param correctly in MIR typeck
Properly intercepts and then annotates the type for a `ConstKind::Param` in the MIR.
This code should probably be cleaned up, it's kinda spaghetti, but no better structure really occurred to me when writing this case.
We could probably gate this behind the feature gate or add a fast path when the args have no free regions if perf is bad.
r? `@BoxyUwU`
Use `rustc_type_ir` directly less in the codebase
cc https://github.com/rust-lang/rust/issues/138449
This is a somewhat opinionated bundle of changes that will make working on https://github.com/rust-lang/rust/issues/138449 more easy, since it cuts out the bulk of the changes that would be necessitated by the lint. Namely:
1. Fold `rustc_middle::ty::fold` and `rustc_middle::ty::visit` into `rustc_middle::ty`. This is because we already reexport some parts of these modules into `rustc_middle::ty`, and there's really no benefit from namespacing away the rest of these modules's functionality given how important folding and visiting is to the type layer.
2. Rename `{Decodable,Encodable}_Generic` to `{Decodable,Encodable}_NoContext`[^why], change it to be "perfect derive" (`synstructure::AddBounds::Fields`), use it throughout `rustc_type_ir` instead of `TyEncodable`/`TyDecodable`.
3. Make `TyEncodable` and `TyDecodable` derives use `::rustc_middle::ty::codec::TyEncoder` (etc) for its generated paths, and move the `rustc_type_ir::codec` module back to `rustc_middle::ty::codec` 🎉.
4. Stop using `rustc_type_ir` in crates that aren't "fundamental" to the type system, namely middle/infer/trait-selection. This amounted mostly to changing imports from `use rustc_type_ir::...` to `use rustc_middle::ty::...`, but also this means that we can't glob import `TyKind::*` since the reexport into `rustc_middle::ty::TyKind` is a type alias. Instead, use the prefixed variants like `ty::Str` everywhere -- IMO this is a good change, since it makes it more regularized with most of the rest of the compiler.
[^why]: `_NoContext` is the name for derive macros with no additional generic bounds and which do "perfect derive" by generating bounds based on field types. See `HashStable_NoContext`.
I'm happy to cut out some of these changes into separate PRs to make landing it a bit easier, though I don't expect to have much trouble with bitrot.
r? lcnr
Do not suggest using `-Zmacro-backtrace` for builtin macros
For macros that are implemented on the compiler, or that are annotated with `rustc_diagnostic_item`, which have arbitrary implementations from the point of view of the user and might as well be intrinsics, we do *not* mention the `-Zmacro-backtrace` flag. This includes `derive`s and standard macros like `panic!` and `format!`.
This PR adds a field to every `Span`'s `ExpnData` stating whether it comes from a builtin macro. This is determined by the macro being annotated with either `#[rustc_builtin_macro]` or `#[rustc_diagnostic_item]`. An alternative to using these attributes that already exist for other uses would be to introduce another attribute like `#[rustc_no_backtrace]` to have finer control on which macros are affected (for example, an error within `vec![]` now doesn't mention the backtrace, but one could make the case that it should). Ideally, instead of carrying this information in the `ExpnData` we'd instead try to query the `DefId` of the macro (that is already stored) to see if it is annotated in some way, but we do not have access to the `TyCtxt` from `rustc_errors`.
r? `@petrochenkov`
Make `Parser::parse_expr_cond` public
This allows usage in rustfmt and rustfmt forks.
I'm using this for custom macro formatting, see 30c83df9e1/src/parse/macros/html.rs (L57)
It would be great if this could be upstreamed so I don't need to rely on a fork.
Fix HIR printing of parameters
HIR pretty printing does the wrong thing for anonymous parameters, and there is no test coverage for it. This PR remedies both of those things.
r? ``@lcnr``
Refactor is_snake_case.
I wondered what the definition of this actually was, and found the original hard to read. I believe this change preserves the original behavior, but is hopefully clearer.
rustc_target: Add target features for LoongArch v1.1
This patch adds new target features for LoongArch v1.1:
* div32
* lam-bh
* lamcas
* ld-seq-sa
* scq
This clarifies the explanation of why this is not allowed and also what to do instead.
Fixes 62071
PS There was suggestion of adding a link to the book. I did not yet do that, but if desired that could be added.
Provide helpful diagnostics for shebang lookalikes
When `[` is not found after a `#!`, a note will be added to the exisiting error
```
error: expected `[`, found `/`
--> src/main.rs:2:3
|
2 | #!/usr/bin/env -S cargo +nightly -Zscript
| ^ expected `[`
|
= note: the token sequence `#!` here looks like the start of a shebang interpreter directive but it is not
= help: if you meant this to be a shebang interpreter directive, move it to the very start of the file
```
Fixes#137249
r? `@fmease`
On Unix the limits can be gargantuan anyway so we're pretty
unlikely to hit them, but might still exceed it.
We consult ARG_MAX here to get an estimate.
EUV: fix place of deref pattern's interior's scrutinee
The place previously used here was that of the temporary holding the reference returned by `Deref::deref` or `DerefMut::deref_mut`. However, since the inner pattern of `deref!(inner)` expects the deref-target type itself, this would ICE when that type was inspected (e.g. by the EUV case for slice patterns). This adds a deref projection to fix that.
Since current in-tree consumers of EUV (upvar inference and clippy) don't care about Rvalues, the place could be simplified to `self.cat_rvalue(pat.hir_id, self.pat_ty_adjusted(subpat)?)` to save some cycles. I personally find EUV to be a bit fragile, so I've opted for pedantic correctness. Maybe a `HACK` comment would suffice though?
Fixes#125059
r? `@compiler-errors`
Visit `PatField` when collecting lint levels
Fixes#138428
Side-note, I vaguely skimmed over the other nodes we could be visiting here and it doesn't *seem* to me that we're missing anything, though I may be mistaken given recent(?) support for attrs in where clauses(??). Can be fixed in a follow-up PR.
fix: remove the check of lld not supporting @response-file
In LLVM v9, lld has supported `@response-file.`
LLVM v9 was released on 2019-09-19.
The check was added back to 2018-03-14 (1.26.0) via 04442af18b.
It has been more than five years, and we ship our own lld regardlessly.
This should be happily removed.
See also:
* <bb12396f91>
* <https://reviews.llvm.org/D63024>
atomic intrinsics: clarify which types are supported and (if applicable) what happens with provenance
The provenance semantics match what Miri implements and what the `AtomicPtr` API expects.
Sync Fuchsia target spec with clang Fuchsia driver
This updates the Fuchsia target spec with the [Clang Fuchsia driver], which picks up a few changes:
* Adds `-z start-stop-visibility=hidden` and `-z rel` to the pre link arguments.
* Adds `--execute-only` and `--fix-cortex-a53-843419` for `aarch64-unknown-fuchsia`.
* Enables the equivalent cpu features for `x86-64-v2` for `x86_64-unknown-fuchsia`, which is our minimum supported x86_64 platform according to [RFC-0073].
try-job: x86_64-fuchsia
[Clang Fuchsia driver]: 8374d42186/clang/lib/Driver/ToolChains/Fuchsia.cpp
[RFC-0073]: https://fuchsia.dev/fuchsia-src/contribute/governance/rfcs/0073_x86_64_platform_requirement
Allow more top-down inlining for single-BB callees
This means that things like `<usize as Step>::forward_unchecked` and `<PartialOrd for f32>::le` will inline even if
we've already done a bunch of inlining to find the calls to them.
Fixes#138136
~~Draft as it's built atop #138135, which adds a mir-opt test that's a nice demonstration of this. To see just this change, look at <48f63e3be5>~~ Rebased to be just the inlining change, as the other existing tests show it great.
Don't `alloca` just to look at a discriminant
Today we're making LLVM do a bunch of extra work when you match on trivial stuff like `Option<bool>` or `ControlFlow<u8>`.
This PR changes that so that simple types like `Option<u32>` or `Result<(), Box<Error>>` can stay as `OperandValue::ScalarPair` and we can still read the discriminant from them, rather than needing to write them into memory to have a `PlaceValue` just to get the discriminant out.
Fixes#137503
Currently (PatKind::Wild` (i.e. `_`) gets turned by
`lower_fn_params_to_names` into an empty identifier, which means it is
printed incorrectly by HIR pretty printing.
And likewise for `lower_fn_params_to_names`, which affects some error
messages.
This commit fixes them. This requires a slight tweak in a couple of
places to continue using parameter numbers in some error messages. And
it improves the output of `tests/ui/typeck/cyclic_type_ice.rs`:
`/* _ */` is a better suggestion than `/* */`.
Fix RISC-V VxWorks LLVM target triples
The targets `riscv32-wrs-vxworks` and `riscv64-wrs-vxworks` uses the plain `$ARCH` LLVM triple, which LLVM normalizes to `$ARCH-unknown-unknown`, we should use `$ARCH-unknown-linux-gnu$ABI` which is consistent with the the other VxWorks targets.
Motivation: To make it easier to verify that [`cc-rs`' conversion from `rustc` to Clang/LLVM triples](https://github.com/rust-lang/cc-rs/issues/1431) is correct.
Alternative: Pass `$ARCH-unknown-none` in the other VxWorks LLVM triples, I don't know anything about VxWorks, so am unsure which is the most correct option.
CC target maintainer `@biabbas.`
r? jieyouxu
