Refactor linker code
This merges `LinkerInfo` into `CrateInfo` as there is no reason to keep them separate. `LinkerInfo::to_linker` is merged into `get_linker` as both have different logic for each linker type and `to_linker` is directly called after `get_linker`. Also contains a couple of small cleanups.
See the individual commits for all changes.
The `large_assignments` lints detects moves over specified limit. The
limit is configured through `move_size_limit = "N"` attribute placed at
the root of a crate. When attribute is absent, the lint is disabled.
Make it possible to enable the lint without making any changes to the
source code, through a new flag `-Zmove-size-limit=N`. For example, to
detect moves exceeding 1023 bytes in a cargo crate, including all
dependencies one could use:
```
$ env RUSTFLAGS=-Zmove-size-limit=1024 cargo build -vv
```
Replace per-target ABI denylist with an allowlist
It makes very little sense to maintain denylists of ABIs when, as far as
non-generic ABIs are concerned, targets usually only support a small
subset of the available ABIs.
This has historically been a cause of bugs such as us allowing use of
the platform-specific ABIs on x86 targets – these in turn would cause
LLVM errors or assertions to fire.
In this PR we got rid of the per-target ABI denylists, and instead compute
which ABIs are supported with a simple match based on, mostly, the
`Target::arch` field. Among other things, this makes it impossible to
forget to consider this problem (in either direction) and forces one to
consider what the ABI support looks like when adding an ABI (rarely)
rather than target (often), which should hopefully also reduce the
cognitive load on both contributors as well as reviewers.
Fixes#57182
Sponsored by: standard.ai
---
## Summary for teams
One significant user-facing change after this PR is that there's now a future compat warning when building…
* `stdcall`, `fastcall`, `thiscall` using code with targets other than 32-bit x86 (i386...i686) or *-windows-*;
* `vectorcall` using code when building for targets other than x86 (either 32 or 64 bit) or *-windows-*.
Previously these ABIs have been accepted much more broadly, even for architectures and targets where this made no sense (e.g. on wasm32) and would fall back to the C ABI. In practice this doesn't seem to be used too widely and the [breakages in crater](https://github.com/rust-lang/rust/pull/86231#issuecomment-866300943) that we see are mostly about Windows-specific code that was missing relevant `cfg`s and just happened to successfully `check` on Linux for one reason or another.
The intention is that this warning becomes a hard error after some time.
It makes very little sense to maintain denylists of ABIs when, as far as
non-generic ABIs are concerned, targets usually only support a small
subset of the available ABIs.
This has historically been a cause of bugs such as us allowing use of
the platform-specific ABIs on x86 targets – these in turn would cause
LLVM errors or assertions to fire.
Fixes#57182
Sponsored by: standard.ai
Query-ify global limit attribute handling
Currently, we read various 'global limits' from inner attributes the crate root (`recursion_limit`, `move_size_limit`, `type_length_limit`, `const_eval_limit`). These limits are then stored in `Sessions`, allowing them to be access from a `TyCtxt` without registering a dependency on the crate root attributes.
This PR moves the calculation of these global limits behind queries, so that we properly track dependencies on crate root attributes. During the setup of macro expansion (before we've created a `TyCtxt`), we need to access the recursion limit, which is now done by directly calling into the code shared by the normal query implementations.
E0716: clarify that equivalent code example is erroneous
In E0716, there is a code block that is equivalent to the erroneous
code example. Especially when viewed with `rustc --explain`, it's
not obvious that it is also erroneous, and some users have been
confused when they try to change their code to match the erroneous
equivalent.
`@rustbot` label +A-diagnostics +D-newcomer-roadblock +T-compiler
Hack: Ignore inference variables in certain queries
Fixes#84841Fixes#86753
Some queries are not built to accept types with inference variables, which can lead to ICEs. These queries probably ought to be converted to canonical form, but as a quick workaround, we can return conservative results in the case that inference variables are found.
We should file a follow-up issue (and update the FIXMEs...) to do the proper refactoring.
cc `@arora-aman`
r? `@oli-obk`
In E0716, there is a code block that is equivalent to the erroneous
code example. Especially when viewed with `rustc --explain`, it's
not obvious that it is also erroneous, and some users have been
confused when they try to change their code to match the erroneous
equivalent.
Support allocation failures when interpreting MIR
This closes#79601 by handling the case where memory allocation fails during MIR interpretation, and translates that failure into an `InterpError`. The error message is "tried to allocate more memory than available to compiler" to make it clear that the memory shortage is happening at compile-time by the compiler itself, and that it is not a runtime issue.
Now that memory allocation can fail, it would be neat if Miri could simulate low-memory devices to make it easy to see how much memory a Rust program needs.
Note that this breaks Miri because it assumes that allocation can never fail.
Warn when `rustdoc::` group is omitted from lint names
When rustdoc lints were first made a tool lint, they gave an unconditional warning when you used the original name:
```
warning: lint `broken_intra_doc_links` has been renamed to `rustdoc::broken_intra_doc_links`
--> $DIR/renamed-lint-still-applies.rs:2:9
|
LL | #![deny(broken_intra_doc_links)]
| ^^^^^^^^^^^^^^^^^^^^^^ help: use the new name: `rustdoc::broken_intra_doc_links`
|
= note: `#[warn(renamed_and_removed_lints)]` on by default
```
That was reverted in https://github.com/rust-lang/rust/pull/83203 because adding `rustdoc::x` lints would cause the code to break on old versions of the compiler (due to https://github.com/rust-lang/rust/issues/66079#issuecomment-788589193, "fixed" in https://github.com/rust-lang/rust/pull/83216 in the sense that you can now opt-in to not breaking on nightly, which is not ideal but `register_tool` is a long way from stabilizing). Since https://github.com/rust-lang/rust/pull/80527 is now on 1.52.0 stable, we can re-enable the warning. For nightly users, they can change immediately and still have their code work on stable; for stable users, they can change their code in 12 weeks and still have it work up to 3 releases back (about 18 weeks). That seems reasonable to me.
r? `@Manishearth` cc `@rust-lang/rustdoc`
Return `EvaluatedToOk` when type in outlives predicate is global
A global type doesn't reference any local regions or types, so it's
guaranteed to outlive any region.
Avoid byte to char position conversions in `is_multiline`
Converting a byte position into a char position is currently linear in
the number of multibyte characters in the source code. Avoid it when
checking if a range spans across lines.
This makes it feasible to compile source files with a large number of
multibyte characters.
deny using default function in impl const Trait
Fixes#79450.
I don't know if my implementation is correct:
- The check is in `rustc_passes::check_const`, should I put it somewhere else instead?
- Is my approach (to checking the impl) optimal? It works for the current tests, but it might have some issues or there might be a better way of doing this.
Fix const-generics ICE related to binding
Fixes#83765, fixes#85848
r? `@jackh726` as you're familiar with `Binding`. I'd like to get some views if the current approach is right path.
Previously, only the fields would be displayed with no indication of the
variant name. If you already knew the enum was univariant, this was ok
but if the enum was univariant because of layout, for example, a
`Result<T, !>` then it could be very confusing which variant was the
active one.
Previously, directly tagged enums had a `variant$` field which would
show the name of the active variant. We now show the variant using a
`[variant]` synthetic item just like we do for niche-layout enums.
Rollup of 7 pull requests
Successful merges:
- #84029 (add `track_path::path` fn for usage in `proc_macro`s)
- #85001 (Merge `sys_common::bytestring` back into `os_str_bytes`)
- #86308 (Docs: clarify that certain intrinsics are not unsafe)
- #86796 (Add a regression test for issue-70703)
- #86803 (Remove & from Command::args calls in documentation)
- #86807 (Fix double import in wasm thread )
- #86813 (Add a help message to `unused_doc_comments` lint)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Add a help message to `unused_doc_comments` lint
Fixes#83492
This adds a help message to suggest a plain comment like the E0658 error. I've yet to come up with the best message about the doc attribute but the current shouldn't harm anything.
I was thinking of recovering in the `doc_comment_between_if_else` case, but I came to the conclusion that it unlikely happened and was an overkill.
add `track_path::path` fn for usage in `proc_macro`s
Adds a way to declare a dependency on external files without including them, to either re-trigger the build of a file as well as covering the use case of including dependencies within the `rustc` invocation, such that tools like `sccache`/`cachepot` are able to handle references to external files which are not included.
Ref #73921
Improve debug symbol names to avoid ambiguity and work better with MSVC's debugger
There are several cases where names of types and functions in the debug info are either ambiguous, or not helpful, such as including ambiguous placeholders (e.g., `{{impl}}`, `{{closure}}` or `dyn _'`) or dropping qualifications (e.g., for dynamic types).
Instead, each debug symbol name should be unique and useful:
* Include disambiguators for anonymous `DefPathDataName` (closures and generators), and unify their formatting when used as a path-qualifier vs item being qualified.
* Qualify the principal trait for dynamic types.
* If there is no principal trait for a dynamic type, emit all other traits instead.
* Respect the `qualified` argument when emitting ref and pointer types.
* For implementations, emit the disambiguator.
* Print const generics when emitting generic parameters or arguments.
Additionally, when targeting MSVC, its debugger treats many command arguments as C++ expressions, even when the argument is defined to be a symbol name. As such names in the debug info need to be more C++-like to be parsed correctly:
* Avoid characters with special meaning (`#`, `[`, `"`, `+`).
* Never start a name with `<` or `{` as this is treated as an operator.
* `>>` is always treated as a right-shift, even when parsing generic arguments (so add a space to avoid this).
* Emit function declarations using C/C++ style syntax (e.g., leading return type).
* Emit arrays as a synthetic `array$<type, size>` type.
* Include a `$` in all synthetic types as this is a legal character for C++, but not Rust (thus we avoid collisions with user types).
Implement printing of stack traces on LLVM segfaults and aborts
Implement #79153
Based on discussion, try to extend the rust_backtrace=1 feature to handle segfault or aborts in the llvm backend
Include terminators in instance size estimate
For example, drop glue generated for struct below, doesn't have any
statements, only terminators. Previously it received an estimate of 0,
the new estimate is 13 (6+5 drop terminators, +1 resume, +1 return).
```rust
struct S {
a: String,
b: String,
c: String,
d: String,
e: String,
f: String,
}
```
Originally reported in https://github.com/rust-lang/rust/issues/69382#issue-569392141
Check the number of generic lifetime and const parameters of intrinsics
This pull request fixes#85855. The current code for type checking intrinsics only checks the number of generic _type_ parameters, but does not check for an incorrect number of lifetime or const parameters, which can cause problems later on, such as the ICE in #85855, where the code thought that it was looking at a type parameter but found a lifetime parameter:
```
error: internal compiler error: compiler/rustc_middle/src/ty/generics.rs:188:18:
expected type parameter, but found another generic parameter
```
The changes in this PR add checks for the number of lifetime and const parameters, expand the scope of `E0094` to also apply to these cases, and improve the error message by properly pluralizing the number of expected generic parameters.
Also an fix issue with tuple type names where we can't cast to them in
natvis (required by the visualizer for `HashMap`) because of
peculiarities with the natvis expression evaluator.
For example, drop glue generated for struct below, doesn't have any
statements, only terminators. Previously it received an estimate of 0,
the new estimate is 13 (6+5 drop terminators, +1 resume, +1 return).
struct S {
a: String,
b: String,
c: String,
d: String,
e: String,
f: String,
}
Originally reported in https://github.com/rust-lang/rust/issues/69382#issue-569392141
Converting a byte position into a char position is currently linear in
the number of multibyte characters in the source code. Avoid it when
checking if a range spans across lines.
This makes it feasible to compile source files with a large number of
multibyte characters.
Fix misleading "impl Trait" error
The kinds can't be compared directly, as types with references are treated as different because the lifetimes aren't bound in ty, but are in expected.
Closes#84160
Add support for leaf function frame pointer elimination
This PR adds ability for the target specifications to specify frame
pointer emission type that's not just “always” or “whatever cg decides”.
In particular there's a new mode that allows omission of the frame
pointer for leaf functions (those that don't call any other functions).
We then set this new mode for Aarch64-based Apple targets.
Fixes#86196
rustc_span: Explicitly handle crates that differ from package names
The sha-1 and md-5 packages contain crates named sha1 and md5,
respectively. This discrepancy makes it somewhat more challenging to
automate detection of unused crates. Explicitly rename the packages to
the names of the crates they contain, to simplify such detection.
Add suggestions for "undefined reference" link errors
This adds a suggestion for "undefined reference to ..." linking errors to install or specify the location to an external library. Since there is no defined error format for linkers, we just check if there was a failure and if that failure contains the string `undefined reference to`. This also makes it impossible to test this, since the output depends on the system linker. The output now looks like:
```
error: linking with `cc` failed: exit status: 1
|
= note: "cc" "-m64" "linking_failure.linking_failure.7rcbfp3g-cgu.0.rcgu.o" "linking_failure.linking_failure.7rcbfp3g-cgu.1.rcgu.o" "linking_failure.linking_failure.7rcbfp3g-cgu.2.rcgu.o" "linking_failure.linking_failure.7rcbfp3g-cgu.3.rcgu.o" "linking_failure.linking_failure.7rcbfp3g-cgu.4.rcgu.o" "linking_failure.linking_failure.7rcbfp3g-cgu.5.rcgu.o" "linking_failure.linking_failure.7rcbfp3g-cgu.6.rcgu.o" "linking_failure.linking_failure.7rcbfp3g-cgu.7.rcgu.o" "linking_failure.linking_failure.7rcbfp3g-cgu.8.rcgu.o" "linking_failure.53u64zklswtfazes.rcgu.o" "-Wl,--as-needed" "-L" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib" "-Wl,--start-group" "-Wl,-Bstatic" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libstd-01ce3ba5c629d02f.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libpanic_unwind-f1f2102409186354.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libminiz_oxide-1e8b6b56a999f838.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libadler-d0e93eb4e14f1d19.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libobject-1d7e39d75d082b43.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libaddr2line-ade42e945045b261.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libgimli-1a65064fccf4ebc1.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libstd_detect-4d699c310fdfe72d.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_demangle-1cafa68a696ec800.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libhashbrown-e9f1c8c4dab2f046.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_std_workspace_alloc-ecc1a743be25c7f7.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libunwind-e074031c4b66b6b6.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcfg_if-9aa6ed9f1d3bfd53.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/liblibc-7862bf96c2250ca0.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/liballoc-f02ce0dc7895b5fd.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_std_workspace_core-3af9c60917570521.rlib" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcore-ca16fc7bb3645684.rlib" "-Wl,--end-group" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib/libcompiler_builtins-d8e1a5b7299604cc.rlib" "-Wl,-Bdynamic" "-lgcc_s" "-lutil" "-lrt" "-lpthread" "-lm" "-ldl" "-lc" "-Wl,--eh-frame-hdr" "-Wl,-znoexecstack" "-L" "/home/smit/rustc-dev/rust/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib" "-o" "linking_failure" "-Wl,--gc-sections" "-pie" "-Wl,-zrelro" "-Wl,-znow" "-nodefaultlibs"
= note: /usr/bin/ld: linking_failure.linking_failure.7rcbfp3g-cgu.3.rcgu.o: in function `linking_failure::main':
linking_failure.7rcbfp3g-cgu.3:(.text._ZN15linking_failure4main17h52b6e3052e444479E+0x3): undefined reference to `doesnt_exist_thiwthwfyl'
clang: error: linker command failed with exit code 1 (use -v to see invocation)
= help: some `extern` functions couldn't be found; you may need to install or specify the path to some dependencies
= note: use the -L flag to specify the library lookup path
= note: use the cargo:rustc-link-search directive to specify the library lookup path with Cargo (see https://doc.rust-lang.org/cargo/reference/build-scripts.html#rustc-link-search)
error: aborting due to previous error
```
Fix ICE when `main` is declared in an `extern` block
Changes in #84401 to implement `imported_main` changed how the crate entry point is found, and a declared `main` in an `extern` block was detected erroneously. This was causing the ICE described in #86110.
This PR adds a check for this case and emits an error instead. Previously a `main` declaration in an `extern` block was not detected as an entry point at all, so emitting an error shouldn't break anything that worked previously. In 1.52.1 stable this is demonstrated, with a `` `main` function not found`` error.
Fixes#86110
Remove unused dependencies from compiler crates
Various compiler crates have dependencies that they don't appear to use. I used some scripting to detect such dependencies, filtered them based on some manual review, and removed those that do indeed appear to be entirely unused.
Test cross-crate usage of `feature(const_trait_impl)`
This PR does two things:
- Fixes metadata not encoded properly for functions in const trait impls.
- Adds tests for using const trait impls cross-crate with the feature gate on the user crate either enabled or disabled.
AFAIK, this means we can now constify some trait impls in the standard library 🎉
See #67792 for the tracking issue, cc `@oli-obk`
Check node kind to avoid ICE in `check_expr_return()`
This PR fixes#86721. The ICE described there is apparently due to a misunderstanding:
e98897e5dc/compiler/rustc_typeck/src/check/expr.rs (L684-L685)
Intuitively, one would think that calling `expect_item()` after `get_parent_item()` should succeed, but as it turns out, `get_parent_item()` can also return foreign, trait, and impl items as well as crates, whereas `expect_item()` specifically expects a `Node::Item`. I have therefore added an extra check to prevent this ICE.
There are several cases where names of types and functions in the debug info are either ambiguous, or not helpful, such as including ambiguous placeholders (e.g., `{{impl}}`, `{{closure}}` or `dyn _'`) or dropping qualifications (e.g., for dynamic types).
Instead, each debug symbol name should be unique and useful:
* Include disambiguators for anonymous `DefPathDataName` (closures and generators), and unify their formatting when used as a path-qualifier vs item being qualified.
* Qualify the principal trait for dynamic types.
* If there is no principal trait for a dynamic type, emit all other traits instead.
* Respect the `qualified` argument when emitting ref and pointer types.
* For implementations, emit the disambiguator.
* Print const generics when emitting generic parameters or arguments.
Additionally, when targeting MSVC, its debugger treats many command arguments as C++ expressions, even when the argument is defined to be a symbol name. As such names in the debug info need to be more C++-like to be parsed correctly:
* Avoid characters with special meaning (`#`, `[`, `"`, `+`).
* Never start a name with `<` or `{` as this is treated as an operator.
* `>>` is always treated as a right-shift, even when parsing generic arguments (so add a space to avoid this).
* Emit function declarations using C/C++ style syntax (e.g., leading return type).
* Emit arrays as a synthetic `array$<type, size>` type.
* Include a `$` in all synthetic types as this is a legal character for C++, but not Rust (thus we avoid collisions with user types).
This PR adds ability for the target specifications to specify frame
pointer emission type that's not just “always” or “whatever cg decides”.
In particular there's a new mode that allows omission of the frame
pointer for leaf functions (those that don't call any other functions).
We then set this new mode for Aarch64-based Apple targets.
Fixes#86196
Introduce -Zprofile-closures to evaluate the impact of 2229
This creates a CSV with name "closure_profile_XXXXX.csv", where the
variable part is the process id of the compiler.
To profile a cargo project you can run one of the following depending on
if you're compiling a library or a binary:
```
cargo +nightly rustc --lib -- -Zprofile-closures
cargo +nightly rustc --bin {binary_name} -- -Zprofile-closures
```
r? `@nikomatsakis`
Only include lint in future_incompatible lint group if not an edition lint
A follow up to #86330 - this only includes lints annotated with `FutureIncompatibleInfo` in the `future_incompatibile` lint group if the future compatibility is not tied to an edition.
We probably want to rename `FutureIncompatibleInfo` to something else since this type is now used to indicate future breakages of all kinds (even those that happen in editions). I'd prefer to do that in a separate PR though.
r? `@nikomatsakis`
Change vtable memory representation to use tcx allocated allocations.
This fixes https://github.com/rust-lang/rust/issues/86324. However i suspect there's more to change before it can land.
r? `@bjorn3`
cc `@rust-lang/miri`
One key observation while going over the closure size profile of rustc
was that we are disjointly capturing one or more fields starting at an
immutable reference.
Disjoint capture over immutable reference doesn't add too much value
because the fields can either be borrowed immutably or copied.
One possible edge case of the optimization is when a fields of a struct
have a longer lifetime than the structure, therefore we can't completely
get rid of all the accesses on top of sharef refs, only the rightmost
one. Here is a possible example:
```rust
struct MyStruct<'a> {
a: &'static A,
b: B,
c: C<'a>,
}
fn foo<'a, 'b>(m: &'a MyStruct<'b>) -> impl FnMut() + 'static {
let c = || drop(&*m.a.field_of_a);
// Here we really do want to capture `*m.a` because that outlives `'static`
// If we capture `m`, then the closure no longer outlives `'static'
// it is constrained to `'a`
}
```
Don't make `rustc_insignificant_dtor` feature gate
This isn't a feature gate, it's an attribute that is feature gated behind the `rustc_attrs` attribute. Closes#85680.
Fix garbled suggestion for missing lifetime specifier
This PR fixes#86667. The suggestion code currently checks whether there is a generic parameter that is not a synthetic `impl Trait` parameter and, if so, suggests to insert a new lifetime `'a` before that generic parameter. However, it does not make sense to insert `'a` in front of an elided lifetime parameter, since these are synthetic as well, which leads to the garbled suggestion in #86667.
Turn non_fmt_panic into a future_incompatible edition lint.
This turns the `non_fmt_panic` lint into a future_incompatible edition lint, so it becomes part of the `rust_2021_compatibility` group. See https://github.com/rust-lang/rust/issues/85894.
This lint produces both warnings about semantical changes (e.g. `panic!("{{")`) and things that will become hard errors (e.g. `panic!("{")`). So I added a `explain_reason: false` that supresses the default "this will become a hard error" or "the semantics will change" message, and instead added a note depending on the situation. (cc `@rylev)`
r? `@nikomatsakis`
Fix `future_prelude_collision` false positive
Fixes#86633
The lint for checking if method resolution of methods named `try_into` will fail in 2021 edition previously would fire on all inherent methods, however for inherent methods that consume `self`, this takes priority over `TryInto::try_into` due to being inherent, while trait method and methods that take `&self` or `&mut self` don't take priority, and thus aren't affected by this false positive.
This fix is rather simple: simply checking if the inherent method doesn't auto-deref or auto-ref (and thus takes `self`) and if so, prevents the lint from firing.
This prevents mistakes where the feature is in the list of incomplete
features but not actually a feature by making the incompleteness a part
of the declaration.
This creates a CSV with name "closure_profile_XXXXX.csv", where the
variable part is the process id of the compiler.
To profile a cargo project you can run one of the following depending on
if you're compiling a library or a binary:
```
cargo +stage1 rustc --lib -- -Zprofile-closures
cargo +stage1 rustc --bin -- -Zprofile-closures
```
Fix type checking of return expressions outside of function bodies
This pull request fixes#86188. The problem is that the current code for type-checking `return` expressions stops if the `return` occurs outside of a function body, while the correct behavior is to continue type-checking the return value expression (otherwise an ICE happens later on because variables declared in the return value expression don't have a type).
Also, I have noticed that it is sometimes not obvious why a `return` is outside of a function body; for instance, in the example from #86188 (which currently causes an ICE):
```rust
fn main() {
[(); return || {
let tx;
}]
}
```
I have changed the error message to also explain why the `return` is considered outside of the function body:
```
error[E0572]: return statement outside of function body
--> ice0.rs:2:10
|
1 | / fn main() {
2 | | [(); return || {
| |__________^
3 | || let tx;
4 | || }]
| ||_____^ the return is part of this body...
5 | | }
| |_- ...not the enclosing function body
```
Reserve prefixed identifiers and literals (RFC 3101)
This PR denies any identifiers immediately followed by one of three tokens `"`, `'` or `#`, which is stricter than the requirements of RFC 3101 but may be necessary according to the discussion at [Zulip].
[Zulip]: https://rust-lang.zulipchat.com/#narrow/stream/268952-edition-2021/topic/reserved.20prefixes/near/238470099
The tracking issue #84599 says we'll add a feature gate named `reserved_prefixes`, but I don't think I can do this because it is impossible for the lexer to know whether a feature is enabled or not. I guess determining the behavior by the edition information should be enough.
Fixes#84599
2229: Capture box completely in move closures
Even if the content from box is used in a sharef-ref context,
we capture the box entirerly.
This is motivated by:
1) We only capture data that is on the stack.
2) Capturing data from within the box might end up moving more data than
the user anticipated.
Closes https://github.com/rust-lang/project-rfc-2229/issues/50
r? `@nikomatsakis`
Fix ICE with `-Zunpretty=hir,typed`
This PR fixes#82328. The `-Zunpretty=hir,typed` pretty-printer maintains an `Option` with type-checking results and sets the `Option` to `Some` when entering a body. However, this leads to an ICE if an expression occurs in a function signature (i.e. outside of a body), such as `128` in
```rust
fn foo(-128..=127: i8) {}
```
This PR fixes the ICE by checking (if necessary) whether the expression's owner has a body, and retrieving type-checking results for that on the fly.
Allow loading of llvm plugins on nightly
Based on a discussion in #82734 / with `@wsmoses.`
Mainly moves [this](0149bc4e7e) behind a -Z flag, so it can only be used on nightly,
as requested by `@nagisa` in https://github.com/rust-lang/rust/issues/82734#issuecomment-835863940
This change allows loading of llvm plugins like Enzyme.
Right now it also requires a shared library LLVM build of rustc for symbol resolution.
```rust
// test.rs
extern { fn __enzyme_autodiff(_: usize, ...) -> f64; }
fn square(x : f64) -> f64 {
return x * x;
}
fn main() {
unsafe {
println!("Hello, world {} {}!", square(3.0), __enzyme_autodiff(square as usize, 3.0));
}
}
```
```
./rustc test.rs -Z llvm-plugins="./LLVMEnzyme-12.so" -C passes="enzyme"
./test
Hello, world 9 6!
```
I will try to figure out how to simplify the usage and get this into stable in a later iteration,
but having this on nightly will already help testing further steps.
Use HTTPS links where possible
While looking at #86583, I wondered how many other (insecure) HTTP links were in `rustc`. This changes most other `http` links to `https`. While most of the links are in comments or documentation, there are a few other HTTP links that are used by CI that are changed to HTTPS.
Notes:
- I didn't change any to or in licences
- Some links don't support HTTPS :(
- Some `http` links were dead, in those cases I upgraded them to their new places (all of which used HTTPS)
Check that `#[cmse_nonsecure_entry]` is applied to a function definition
This PR fixes#83475. The compiler currently neglects to check whether `#[cmse_nonsecure_entry]` is applied to a function (and not, say, a struct) definition, leading to an ICE later on when the type checker attempts to retrieve the function signature. I have fixed this problem by adding an appropriate check to the `check_attr` pass, so that an error is reported instead of an ICE.
Rollup of 5 pull requests
Successful merges:
- #86330 (Change how edition based future compatibility warnings are handled)
- #86513 (Rustdoc: Do not list impl when trait has doc(hidden))
- #86592 (Use `#[non_exhaustive]` where appropriate)
- #86608 (chore(rustdoc): remove unused members of RenderType)
- #86624 (Update compiler-builtins)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Use `#[non_exhaustive]` where appropriate
Due to the std/alloc split, it is not possible to make `alloc::collections::TryReserveError::AllocError` non-exhaustive without having an unstable, doc-hidden method to construct (which negates the benefits from `#[non_exhaustive]`).
`@rustbot` label +C-cleanup +T-libs +S-waiting-on-review
rustc_data_structures has a dependency on crossbeam-utils but never uses
it. It appears to have originally had this dependency in order to set
the "nightly" feature; however, its other dependencies use a different
version of crossbeam-utils, so this doesn't actually affect anything.
Furthermore, in current crossbeam-utils, the "nightly" feature has
become a no-op.
The sha-1 and md-5 packages contain crates named sha1 and md5,
respectively. This discrepancy makes it somewhat more challenging to
automate detection of unused crates. Explicitly rename the packages to
the names of the crates they contain, to simplify such detection.
Don't lint :pat when re-parsing a macro from another crate.
`compile_macro` is used both when compiling the original definition in the crate that defines it, and to compile the macro when loading it when compiling a crate that uses it. We should only emit lints in the first case.
This adds a `is_definition: bool` to pass this information in, so we don't warn about things that only concern the definition site.
Fixes#86567
Even if the content from box is used in a sharef-ref context,
we capture the box entirerly.
This is motivated by:
1) We only capture data that is on the stack.
2) Capturing data from within the box might end up moving more data than
the user anticipated.
Fix use placement for suggestions near main.
This fixes an edge case for the suggestion to add a `use`. When running with `--test`, the `main` function will be annotated with an `#[allow(dead_code)]` attribute. The `UsePlacementFinder` would end up using the dummy span of that synthetic attribute. If there are top-level inner attributes, this would place the `use` in the wrong position. The solution here is to ignore attributes with dummy spans.
In the process of working on this, I discovered that the `use_suggestion_placement` test was broken. `UsePlacementFinder` is unaware of active attributes. Attributes like `#[derive]` don't exist in the AST since they are removed. Fixing that is difficult, since the AST does not retain enough information. I considered trying to place the `use` towards the top of the module after any `extern crate` items, but I couldn't find a way to get a span for the start of a module block (the `mod` span starts at the `mod` keyword, and it seems tricky to find the spot just after the opening bracket and past inner attributes). For now, I just put some comments about the issue. This appears to have been a known issue in #44215 where the test for it was introduced, and the fix seemed to be deferred to later.
Due to the std/alloc split, it is not possible to make
`alloc::collections::TryReserveError::AllocError` non-exhaustive without
having an unstable, doc-hidden method to construct (which negates the
benefits from `#[non_exhaustive]`.
Permit zero non-zero-field on transparent types
Fixes#77841
This makes the transparent fields meet the below:
> * A `repr(transparent)` type `T` must meet the following rules:
> * It may have any number of 1-ZST fields
> * In addition, it may have at most one other field of type U
r? `@nikomatsakis`
Support lowercase error codes in `--explain`
This enables `rustc --explain` to accept a lowercase error code. Thus, for instance, `rustc --explain e0573` would be valid after this change, where before a user would have needed to do `rustc --explain E0573`. Although the upper case form of an error code is canonical, the user may prefer the easier-to-type lowercase form, and there's nothing to be gained by forcing them to type the upper case version.
Resolves#86518.
Error code cleanup and enforce checks
Fixes#86097.
It now checks if an error code is unused, and if so, will report an error if the error code wasn't commented out in the `error_codes.rs` file. It also checks that the constant used in the tidy check is up-to-date.
r? `@Mark-Simulacrum`
Check whether the closure's owner is an ADT in thir-unsafeck
This pull request fixes#85871. The code in `rustc_mir_build/src/check_unsafety.rs` incorrectly assumes that a closure's owner always has a body, but only functions, closures, and constants have bodies, whereas a closure can also appear inside a struct or enum:
```rust
struct S {
arr: [(); match || 1 { _ => 42 }]
}
enum E {
A([(); { || 1; 42 }])
}
```
This pull request fixes the resulting ICE by checking whether the closure's owner is an ADT and only deferring to `thir_check_unsafety(owner)` if it isn't.
Better errors for Debug and Display traits
Currently, if someone tries to pass value that does not implement `Debug` or `Display` to a formatting macro, they get a very verbose and confusing error message. This PR changes the error messages for missing `Debug` and `Display` impls to be less overwhelming in this case, as suggested by #85844. I was a little less aggressive in changing the error message than that issue proposed. Still, this implementation would be enough to reduce the number of messages to be much more manageable.
After this PR, information on the cause of an error involving a `Debug` or `Display` implementation would suppressed if the requirement originated within a standard library macro. My reasoning was that errors originating from within a macro are confusing when they mention details that the programmer can't see, and this is particularly problematic for `Debug` and `Display`, which are most often used via macros. It is possible that either a broader or a narrower criterion would be better. I'm quite open to any feedback.
Fixes#85844.
Rollup of 6 pull requests
Successful merges:
- #86223 (Specify the kind of the item for E0121)
- #86521 (Add comments around code where ordering is important due for panic-safety)
- #86523 (Improvements to intra-doc link macro disambiguators)
- #86542 (Line numbers aligned with content)
- #86549 (Add destructuring example of E0508)
- #86557 (Update books)
Failed merges:
- #86548 (Fix crate filter search reset)
r? `@ghost`
`@rustbot` modify labels: rollup
Re-add support for parsing (and pretty-printing) inner-attributes in match body
Re-add support for parsing (and pretty-printing) inner-attributes within body of a `match`.
In other words, we can do `match EXPR { #![inner_attr] ARM_1 ARM_2 ... }` again.
I believe this unbreaks the only four crates that crater flagged as broken by PR #83312.
(I am putting this up so that the lang-team can check it out and decide whether it changes their mind about what to do regarding PR #83312.)
Fix emit path hashing
With `--emit KIND=PATH`, the PATH should not affect hashes used for dependency tracking. It does not with other ways of specifying output paths (`-o` or `--out-dir`).
Also updates `rustc -Zls` to print more info about crates, which is used here to implement a `run-make` test.
It seems there was already a test explicitly checking that `OutputTypes` hash *is* affected by the path. I think this behaviour is wrong, so I updated the test.
Disambiguate between SourceFiles from different crates even if they have the same path
This PR fixes an ICE that can occur when the compiler encounters a source file that is part of both the local crate and an upstream crate:
1. While importing source files from an upstream crate the compiler creates a `SourceFile` entry for `foo.rs` in the `SourceMap`. Since this is an imported source file its `src` field is `None`.
2. At a later point the parser encounters `foo.rs` again. It tells the `SourceMap` to load the file but because we already have an entry for `foo.rs` the `SourceMap` will return the existing version with `src == None`.
3. The parser proceeds under the assumption that `src.is_some()` and panics when actually trying to use the file's contents.
This PR fixes the issue by adding the source file's associated `CrateNum` to the `SourceMap`'s interning key. As a consequence the two instances of the file will each have a separate entry in the `SourceMap`. They just happen to share the same file path. This approach seemed less problematic to me than trying to mutate the `SourceFile` after it had already been created.
Another, more involved, approach might be to merge the `src` and the `external_src` field.
Fixes#85955
Fix `unused_unsafe` around `await`
Enables `unused_unsafe` lint for `unsafe { future.await }`.
The existing test for this is `unsafe { println!() }`, so I assume that `println!` used to contain compiler-generated unsafe but this is no longer true, and so the existing test is broken. I replaced the test with `unsafe { ...await }`. I believe `await` is currently the only instance of compiler-generated unsafe.
Reverts some parts of #85421, but the issue predates that PR.
Add `future_prelude_collision` lint
Implements #84594. (RFC rust-lang/rfcs#3114 ([rendered](https://github.com/rust-lang/rfcs/blob/master/text/3114-prelude-2021.md))) Not entirely complete but wanted to have my progress decently available while I finish off the last little bits.
Things left to implement:
* [x] UI tests for lints
* [x] Only emit lint for 2015 and 2018 editions
* [ ] Lint name/message bikeshedding
* [x] Implement for `FromIterator` (from best I can tell, the current approach as mentioned from [this comment](https://github.com/rust-lang/rust/issues/84594#issuecomment-847288288) won't work due to `FromIterator` instances not using dot-call syntax, but if I'm correct about this then that would also need to be fixed for `TryFrom`/`TryInto`)*
* [x] Add to `rust-2021-migration` group? (See #85512) (added to `rust-2021-compatibility` group)
* [ ] Link to edition guide in lint docs
*edit: looked into it, `lookup_method` will also not be hit for `TryFrom`/`TryInto` for non-dotcall syntax. If anyone who is more familiar with typecheck knows the equivalent for looking up associated functions, feel free to chime in.
Rollup of 11 pull requests
Successful merges:
- #85054 (Revert SGX inline asm syntax)
- #85182 (Move `available_concurrency` implementation to `sys`)
- #86037 (Add `io::Cursor::{remaining, remaining_slice, is_empty}`)
- #86114 (Reopen#79692 (Format symbols under shared frames))
- #86297 (Allow to pass arguments to rustdoc-gui tool)
- #86334 (Resolve type aliases to the type they point to in intra-doc links)
- #86367 (Fix comment about rustc_inherit_overflow_checks in abs().)
- #86381 (Add regression test for issue #39161)
- #86387 (Remove `#[allow(unused_lifetimes)]` which is now unnecessary)
- #86398 (Add regression test for issue #54685)
- #86493 (Say "this enum variant takes"/"this struct takes" instead of "this function takes")
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Say "this enum variant takes"/"this struct takes" instead of "this function takes"
This makes error messages for functions with incorrect argument counts adapt if they refer to a struct or enum variant:
```
error[E0061]: this enum variant takes 1 argument but 0 arguments were supplied
--> $DIR/struct-enum-wrong-args.rs:7:13
|
LL | let _ = Ok();
| ^^-- supplied 0 arguments
| |
| expected 1 argument
error[E0061]: this struct takes 1 argument but 0 arguments were supplied
--> $DIR/struct-enum-wrong-args.rs:8:13
|
LL | let _ = Wrapper();
| ^^^^^^^-- supplied 0 arguments
| |
| expected 1 argument
```
Fixes#86481.
Add MIR pass to lower call to `core::slice::len` into `Len` operand
During some larger experiment with range analysis I've found that code like `let l = slice.len()` produces different MIR then one found in bound checks. This optimization pass replaces terminators that are calls to `core::slice::len` with just a MIR operand and Goto terminator.
It uses some heuristics to remove the outer borrow that is made to call `core::slice::len`, but I assume it can be eliminated, just didn't find how.
Would like to express my gratitude to `@oli-obk` who helped me a lot on Zullip
Emit warnings for unused fields in custom targets.
Add a warning which lists any fields in a custom target `json` file that aren't used. Currently unrecognized fields are ignored so, for example, a typo in the `json` will silently produce a target which isn't the one intended.
Do not emit alloca for ZST locals with multiple assignments
This extends 35566bfd7d to additionally stop emitting unnecessary allocas for zero sized locals that are assigned multiple times.
When rebuilding the standard library with `-Zbuild-std` this reduces the number of locals that require an allocation from 62315 to 61767.
Replace some `std::iter::repeat` with `str::repeat`
I noticed that there were some instances where `std::iter::repeat` would be used to repeat a string or a char to take a specific count of it and then collect it into a `String` when `str::repeat` is actually much faster and better for that.
See also: https://github.com/rust-lang/rust-clippy/issues/7260.
Add pattern walking support to THIR walker
Suggested in https://github.com/rust-lang/rust/pull/85263#issuecomment-861906730, this splits off the support for pattern walking in THIR from #85263. This has no observable effect on THIR unsafety checking, since it is not currently possible to trigger unsafety from the THIR checker using the additional patterns or constants that are now walked. THIR patterns are walked in source code order.
r? `@LeSeulArtichaut`
Fix ICE with `#[repr(simd)]` on enum
This pull request fixes#83505. `#[repr(simd)]` may only be applied to structs, which correctly causes `E0517` for the example given in #83505, but the compiler attempts to recover from this error, which leads to an ICE later, when `.non_enum_variant()` is called on the `AdtDef`. I have added a check that prevents this from happening.
make UB during CTFE a hard error
This is a next step for https://github.com/rust-lang/rust/issues/71800. `const_err` has been a future-incompatibility lint for 4 months now since https://github.com/rust-lang/rust/pull/80394 (and err-by-default for many years before that), so I think we could try making it a proper hard error at least in some situations.
I didn't yet adjust the tests, since I first want to gauge the fall-out via crater.
Cc `@rust-lang/wg-const-eval`
Remove some last remants of {push,pop}_unsafe!
These macros have already been removed, but there was still some code handling these macros. That code is now removed.
Provide option for specifying the profiler runtime
Currently, if `-Zinstrument-coverage` is enabled, the target is linked
against the `library/profiler_builtins` crate (which pulls in LLVM's
compiler-rt runtime).
This option enables backends to specify an alternative runtime crate for
handling injected instrumentation calls.
Use `AttrVec` for `Arm`, `FieldDef`, and `Variant`
Uses `AttrVec` for `Arm`, `FieldDef`, and `Variant`, i.e., where the size of the vector can be empty often.
Skips `Crate` and `Item` because I think they may have the attributes on common cases and need more work outside of `rustc_ast` (e.g. rustc_expand needs a lot of tweaks). But if it's reasonable to change, I'm happy to do so.
Fixes#77662
Fix ICE when using `#[doc(keyword = "...")]` on non-items
This pull request fixes#83512. The code for checking attributes calls `expect_item()` when it shouldn't, thus causing an ICE. I have implemented a proper check for the node kind, so that an error is reported instead of the ICE.
Make `s` pre-interned
Now we should be able to pre-intern `s` as the test `ui/lint/rfc-2457-non-ascii-idents/lint-confusable-idents.rs` no longer fails.
Prefer `partition_point` to look up assoc items
Since we now have `partition_point` (instead of `equal_range`), I think it's worth trying to use it instead of manually finding it.
`partition_point` uses `binary_search_by` internally (#85406) and its performance has been improved (#74024), so I guess this will make a performance difference.
Use better error message for hard errors in CTFE
I noticed this while working on #86255: currently the same message is used for hard errors and soft errors in CTFE. This changes the error messages to make hard errors use a message that indicates the reality of the situation correctly, since usage of the constant is never allowed when there was a hard error evaluating it. This doesn't affect the behaviour of these error messages, only the content.
This changes the error logic to check if the error should be hard or soft where it is generated, instead of where it is emitted, to allow this distinction in error messages.
Replace parent substs of associated types with inference vars in borrow checker
Fixes https://github.com/rust-lang/rust/issues/83190
Fixes https://github.com/rust-lang/rust/issues/78450
When we normalize an associated type that refers to an opaque type, it can happen that the substs of the associated type do not occur in the projection (they are parent substs). We previously didn't replace those substs with inference vars, which left a concrete region after all regions should have already been replaced with inference vars and triggered a `delay_span_bug`. After we normalize the opaque type, we now try to replace any remaining concrete regions with inference vars.
Handle C-variadic arguments properly when reporting region errors
This pull request fixes#86053. The issue is that for a C-variadic function
```rust
#![feature(c_variadic)]
unsafe extern "C" fn foo(_: (), ...) {}
```
`foo`'s signature will contain only the first parameter (and have `c_variadic` set to `true`), whereas its body has a second argument (a `hir::Pat` for the `...`).
The code for reporting region errors iterates over the body's parameters and tries to fetch the corresponding parameter from the signature; this causes an out-of-bounds ICE for the `...` (though not in the example above, because there are no region errors to report).
I have simply restricted the iteration over the body parameters to exclude `...`, which is fine because `...` cannot cause a region error.
Stop returning a value from `report_assert_as_lint`
This function only ever returns `None`. Make that explicity by not returning a value at all.
`@rustbot` modify labels +C-cleanup +T-compiler
Fix span calculation in format strings
This pull request fixes#86085. The ICE described there is due to an error in the span calculation inside format strings, if the format string is the result of a macro invocation:
```rust
fn main() {
format!(concat!("abc}"));
}
```
currently produces:
```
error: invalid format string: unmatched `}` found
--> test.rs:2:17
|
2 | format!(concat!("abc}"));
| ^ unmatched `}` in format string
```
which is obviously incorrect. This happens because the span of the entire `concat!()` is combined with the _relative_ location of the unmatched `` `}` `` in the _result_ of the macro invocation (i.e. 4).
In #86085, this has led to a span that starts or ends in the middle of a multibyte character, but the root cause was the same. This pull request fixes the problem.
Box `thir::ExprKind::Adt` for performance
`Adt` is the biggest variant in the enum and probably isn't used very often compared to the other expr kinds, so boxing it should be beneficial for performance. We need a perf test to be sure.
Refactor vtable codegen
This refactor the codegen of vtables of miri interpreter, llvm, cranelift codegen backends.
This is preparation for the implementation of trait upcasting feature. cc #65991
Note that aside from code reorganization, there's an internal behavior change here that now InstanceDef::Virtual's index now include the three metadata slots, and now the first method is with index 3.
cc `@RalfJung` `@bjorn3`
Currently the same message is used for hard errors and soft errors. This
makes hard errors use a message that indicates the reality of the
situation correctly, since usage of the constant is never allowed when
there was a hard error evaluating it.
Improve CTFE UB validation error messages
As mentioned in https://github.com/rust-lang/rust/pull/86245#discussion_r650494012 this PR slightly improves the formatting of validation errors, to move the path to the error prefix.
From:
`type validation failed: encountered invalid vtable: size is bigger than largest supported object at .0`
To:
`type validation failed at .0: encountered invalid vtable: size is bigger than largest supported object`.
Fix force-warns to allow dashes.
The `--force-warns` flag was not allowing lint names with dashes, only supporting underscores. This changes it to allow dashes to match the behavior of the A/W/D/F flags.
Fix ICEs on invalid vtable size/alignment const UB errors
The invalid vtable size/alignment errors from `InterpCx::read_size_and_align_from_vtable` were "freeform const UB errors", causing ICEs when reaching validation. This PR turns them into const UB hard errors to catch them during validation and avoid that.
Fixes#86193
r? `@RalfJung`
(It seemed cleaner to have 2 variants but they can be merged into one variant with a message payload if you prefer that ?)
Do not suggest to add type annotations for unnameable types
Consider this example:
```rust
const A = || 42;
struct S<T> { t: T }
const B: _ = S { t: || 42 };
```
This currently produces the following output:
```
error: missing type for `const` item
--> src/lib.rs:1:7
|
1 | const A = || 42;
| ^ help: provide a type for the item: `A: [closure@src/lib.rs:1:11: 1:16]`
error[E0121]: the type placeholder `_` is not allowed within types on item signatures
--> src/lib.rs:4:10
|
4 | const B: _ = S { t: || 42 };
| ^
| |
| not allowed in type signatures
| help: replace `_` with the correct type: `S<[closure@src/lib.rs:4:21: 4:26]>`
error: aborting due to 2 previous errors
```
However, these suggestions are obviously useless, because the suggested types cannot be written down. With my changes, the suggestion is replaced with a note, because there is no simple fix:
```
error: missing type for `const` item
--> test.rs:1:7
|
1 | const A = || 42;
| ^
|
note: however, the inferred type `[closure@test.rs:1:11: 1:16]` cannot be named
--> test.rs:1:11
|
1 | const A = || 42;
| ^^^^^
error[E0121]: the type placeholder `_` is not allowed within types on item signatures
--> test.rs:4:10
|
4 | const B: _ = S { t: || 42 };
| ^ not allowed in type signatures
|
note: however, the inferred type `S<[closure@test.rs:4:21: 4:26]>` cannot be named
--> test.rs:4:14
|
4 | const B: _ = S { t: || 42 };
| ^^^^^^^^^^^^^^
error: aborting due to 2 previous errors
```
Hash DefId in rustc_span.
This is mostly just moving code around. Changes are simplifications of unneeded callbacks from rustc_span to rustc_middle.
r? `@petrochenkov`
Make `relate_type_and_mut` public
#85343 improved diagnostics around `Relate` impls but made `relate_type_and_mut` private, which was accessible as `relate` previously. This makes it public so that we can use it on rust-semverver.
r? ```@Aaron1011```
Detect incorrect vtable alignment during const eval
This PR fixes#86132 by detecting invalid alignment values for trait objects in the interpreter, and emitting an error about this conversion failure, to avoid the ICE.
I've noticed that the error emitted at a50d72158e/compiler/rustc_mir/src/interpret/traits.rs (L163-L166) doesn't seem to be present in the const-ub tests, so I've tried adding a test that triggers both of these cases: one for the invalid size, and another for the invalid alignment that #86132 tracks (I have found different magic values triggering different `Align::from_bytes` errors than the "power of 2" one, if need be).
However, when doing that, I *cannot* for the life of me figure out the correct incantation to make these 2 errors trigger with the "it is undefined behavior to use this value" message rather than the "any use of this value will cause an error" lint.
I've tried Oli's suggestions of different values, tuples and arrays, using the transparent wrapper trick from the other tests and I was only able to trigger the regular const-ub errors about the size of the vtable, or that the drop pointer was invalid. Maybe these "type validation failed" errors happen before this part of the interpreter is reached and there just needs some magic incorrect values to bypass them, I don't know.
Since this fixes an ICE, and if the constants are indeed used, these 2 tests will turn into a hard error, I thought I'd open the PR anyways. And if ```@RalfJung``` you know of a way I could manage that (if you think that these tests are worth checking that the `throw_ub_format!` does indeed create const-ub errors as we expect) I'd be grateful.
For that reason, r? ```@RalfJung``` and cc ```@oli-obk.```
Do not suggest ampmut if rhs is already mutable
Removes invalid suggestion in #85765, although it should highlight the user type instead of the local variable.
Looking at the comments of this line:
84b1005bfd/compiler/rustc_mir_build/src/build/matches/mod.rs (L2107)
It was intentionally set to `None`, causing it to highlight the local variable instead. I am not sure if I will be able to fix it.
Fixes#85765
Fix some diagnostic issues with const_generics_defaults feature gate
This PR makes a few changes:
- print out const param defaults in "lifetime ordering" errors rather than discarding them
- update `is_simple_text` to account for const params when checking if a type has no generics, this was causing a note to be failed to add to an error message
- fixes some diagnostic wording that incorrectly said there was ordering restrictions between type/const params despite the `const_generics_defaults` feature gate is active
Don't use a generator for BoxedResolver
The generator is non-trivial and requires unsafe code anyway. Using regular unsafe code without a generator is much easier to follow.
Based on #85810 as it touches rustc_interface too.
Suggest a trailing comma if a 1-tuple is expected and a parenthesized expression is found
This pull request fixes#86100. The following code:
```rust
fn main() {
let t: (i32,) = (1);
}
```
currently produces:
```
warning: unnecessary parentheses around assigned value
--> test.rs:2:21
|
2 | let t: (i32,) = (1);
| ^^^ help: remove these parentheses
|
= note: `#[warn(unused_parens)]` on by default
error[E0308]: mismatched types
--> test.rs:2:21
|
2 | let t: (i32,) = (1);
| ------ ^^^ expected tuple, found integer
| |
| expected due to this
|
= note: expected tuple `(i32,)`
found type `{integer}`
error: aborting due to previous error; 1 warning emitted
```
With my changes, I get the same warning and the following error:
```
error[E0308]: mismatched types
--> test.rs:2:21
|
2 | let t: (i32,) = (1);
| ------ ^^^ expected tuple, found integer
| |
| expected due to this
|
= note: expected tuple `(i32,)`
found type `{integer}`
help: use a trailing comma to create a tuple with one element
|
2 | let t: (i32,) = (1,);
| ^^^^
```
i.e. I have added a suggestion to add a trailing comma to create a 1-tuple. This suggestion is only issued if a 1-tuple is expected and the expression (`(1)` in the example above) is surrounded by parentheses and does not already have a tuple type. In this situation, I'd say that it is pretty likely that the user meant to create a tuple.
std: Stabilize wasm simd intrinsics
This commit performs two changes to stabilize Rust support for
WebAssembly simd intrinsics:
* The stdarch submodule is updated to pull in rust-lang/stdarch#1179.
* The `wasm_target_feature` feature gate requirement for the `simd128`
feature has been removed, stabilizing the name `simd128`.
This should conclude the FCP started on #74372 and...
Closes#74372
This commit performs two changes to stabilize Rust support for
WebAssembly simd intrinsics:
* The stdarch submodule is updated to pull in rust-lang/stdarch#1179.
* The `wasm_target_feature` feature gate requirement for the `simd128`
feature has been removed, stabilizing the name `simd128`.
This should conclude the FCP started on #74372 and...
Closes#74372
MVP for using rust-lld as part of cc
Will fix#71519. I need to figure out how to write a test showing that lld is used instead of whatever linker cc normally uses. When I manually run rustc using `echo 'fn main() {}' | RUSTC_LOG=rustc_codegen_ssa:🔙:link=debug ./rustc -Clinker-flavor=gcc-lld --crate-type bin -Clink-arg=-Wl,-v` (thanks to bjorn3 on Zulip), I can see that lld is used, but I'm not sure how to inspect that output in a test.
Disable the machine outliner by default
This addresses a codegen-issue that needs to be fixed upstream in LLVM.
While we wait for the fix, we can disable it.
Verified manually that the outliner is no longer run when
`-Copt-level=z` is specified, and also that you can override this with
`-Cllvm-args=-enable-machine-outliner` if you need it anyway.
A regression test is not really feasible in this instance, given that we
do not have any minimal reproducers.
Fixes#85351
cc `@pnkfelix`
Use preorder traversal when checking for SSA locals
Traverse blocks in topological sort of dominance partial order, to ensure that
local analyzer correctly identifies locals that are already in static single
assignment form, while avoiding dependency on implicit numeric order of blocks.
When rebuilding the standard library, this change reduces the number of locals
that require an alloca from 62452 to 62348.
ignore test if rust-lld not found
create ld -> rust-lld symlink at build time instead of run time
for testing in ci
copy instead of symlinking
remove linux check
test for linker, suggestions from bjorn3
fix overly restrictive lld matcher
use -Zgcc-ld flag instead of -Clinker-flavor
refactor code adding lld to gcc path
revert ci changes
suggestions from petrochenkov
rename gcc_ld to gcc-ld in dirs
Rollup of 7 pull requests
Successful merges:
- #82037 (Make symbols stripping work on MacOS X)
- #84687 (Multiple improvements to RwLocks)
- #85997 (rustdoc: Print a warning if the diff when comparing to old nightlies is empty)
- #86051 (Updated code examples and wording in move keyword documentation )
- #86111 (fix off by one in `std::iter::Iterator` documentation)
- #86113 (build doctests with lld if use-lld = true)
- #86175 (update Miri)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
As reported in the stabilization issue, MacOS' linker doesn't support the `-s` and `-S` flags to strip symbols anymore. However, the os ships a separated tool to perform these operations.
This change allows the compiler to use that tool after a target has been compiled to strip symbols.
For rationale, see: https://github.com/rust-lang/rust/issues/72110#issuecomment-641169818
For option selection, see: https://www.unix.com/man-page/osx/1/strip/
Signed-off-by: David Calavera <david.calavera@gmail.com>
Peephole optimize `x == false` and `x != true`
This adds peephole optimizations to make `x == false`, `false == x`, `x != true`, and `true != x` get optimized to `!x` in the `instcombine` MIR pass. That pass currently handles `x == true` -> `x` already.
Include macro name in 'local ambiguity' error
Currently, we only point at the span of the macro argument. When the
macro call is itself generated by another macro, this can make it
difficult or impossible to determine which macro is responsible for
producing the error.
Enable rustdoc to document safe wasm intrinsics
This commit fixes an issue not found during #84988 where rustdoc is used
to document cross-platform intrinsics but it was requiring that
functions which use `#[target_feature]` are `unsafe` erroneously, even
if they're WebAssembly specific. Rustdoc today, for example, already has
a special case where it enables annotations like
`#[target_feature(enable = "simd128")]` on platforms other than
WebAssembly. The purpose of this commit is to relax the "require all
`#[target_feature]` functions are `unsafe`" requirement for all targets
whenever rustdoc is running, enabling all targets to fully document
other targets, such as WebAssembly, where intrinsics functions aren't
always `unsafe`.
Comment out unused error codes and add description for E0316
I have added an extended description of `E0316` and commented out a bunch of unused error codes to make clear the fact that they are no longer in use. You can check for yourself with
```shell
for ec in \
E0314 E0315 E0473 E0474 E0475 E0479 E0480 E0481 \
E0483 E0484 E0485 E0486 E0487 E0488 E0489
do
if [ ! -z "`grep -r $ec compiler/* --exclude-dir=rustc_error_codes`" ]
then
echo $ec
false
fi
done
```
i.e. these error codes appear nowhere in the compiler code and thus cannot be emitted.
r? ```@GuillaumeGomez```
Currently, we only point at the span of the macro argument. When the
macro call is itself generated by another macro, this can make it
difficult or impossible to determine which macro is responsible for
producing the error.
This commit fixes an issue not found during #84988 where rustdoc is used
to document cross-platform intrinsics but it was requiring that
functions which use `#[target_feature]` are `unsafe` erroneously, even
if they're WebAssembly specific. Rustdoc today, for example, already has
a special case where it enables annotations like
`#[target_feature(enable = "simd128")]` on platforms other than
WebAssembly. The purpose of this commit is to relax the "require all
`#[target_feature]` functions are `unsafe`" requirement for all targets
whenever rustdoc is running, enabling all targets to fully document
other targets, such as WebAssembly, where intrinsics functions aren't
always `unsafe`.
Remove the install prefix from the rpath set when using -Crpath
It was broken anyway for rustup installs and nobody seems to have noticed.
Fixes https://github.com/rust-lang/rust/issues/82392
Unify duplicate linker_and_flavor methods in rustc_codegen_{cranelift,ssa}.
The two methods were exactly the same so this removes the cranelift copy. This will help make sure both they don't get out of sync.
Fix ICE during type layout when there's a `[type error]`
Fixes#84108.
Based on estebank's [comment], except I used `delay_span_bug` because it
should work in more cases, and I think it expresses its intent more
clearly.
r? `@estebank`
[comment]: https://github.com/rust-lang/rust/issues/84108#issuecomment-818916848
Driver improvements
This PR contains a couple of cleanups for the driver and a few small improvements for the custom codegen backend interface. It also implements `--version` and `-Cpasses=list` support for custom codegen backends.
Fix corrected example in E0759.md
This pull request fixes#86061, which was probably caused by a copy-paste error, where the supposedly corrected code example was also marked with `compile_fail`. Thus, the fact that the "correct" example actually _isn't_ correct was not caught by the doc-tests. This pull request removes the incorrect `compile_fail` annotation and fixes the example.
r? ``@GuillaumeGomez``
Fix two ICEs in the parser
This pull request fixes#84104 and fixes#84148. The latter is caused by an invalid `assert_ne!()` in the parser, which I have simply removed because the error is then caught in another part of the parser.
#84104 is somewhat more subtle and has to do with a suggestion to remove extraneous `<` characters; for instance:
```rust
fn main() {
foo::<Ty<<<i32>();
}
```
currently leads to
```
error: unmatched angle brackets
--> unmatched-langle.rs:2:10
|
2 | foo::<Ty<<<i32>();
| ^^^ help: remove extra angle brackets
```
which is obviously wrong and stems from the fact that the code for issuing the above suggestion does not consider the possibility that there might be other tokens in between the opening angle brackets. In #84104, this has led to a span being generated that ends in the middle of a multi-byte character (because the code issuing the suggestion thought that it was only skipping over `<`, which are single-byte), causing an ICE.
Add variance-related information to lifetime error messages
This PR adds a basic framework for displaying variance-related information in error messages. For example:
```
error: lifetime may not live long enough
--> $DIR/type-check-pointer-comparisons.rs:12:5
|
LL | fn compare_mut<'a, 'b>(x: *mut &'a i32, y: *mut &'b i32) {
| -- -- lifetime `'b` defined here
| |
| lifetime `'a` defined here
LL | x == y;
| ^ requires that `'a` must outlive `'b`
|
= help: consider adding the following bound: `'a: 'b`
= note: requirement occurs because of a mutable pointer to &i32
= note: mutable pointers are invariant over their type parameter
= help: see <https://doc.rust-lang.org/nomicon/subtyping.html> for more information about variance
```
The last three lines are new.
This is accomplished by adding a new struct `VarianceDiagInfo`, and passing it along through the various relation methods. When relating types that change the variance (e.g. `&mut T` or `*mut T`), we pass a more specific `VarianceDiagInfo` storing information about the cause of the variance change. When an error, we use the `VarianceDiagInfo` to add additional information to the error message.
This PR doesn't change any variance-related computation or behavior - only diagnostic messages. Therefore, the implementation is quite incomplete - more detailed error messages can be filled in in subsequent PRs.
Limitations:
* We only attempt to deal with invariance - since it's at the bottom of the 'variance lattice', our variance will never change again after it becomes invariant. Handling contravariance would be trickier, since we can change between contravariance and covariance multiple times (e.g. `fn(fn(&'static u8))`). Since contravariance (AFAIK) is only used for function arguments, we can probably get away without a very fancy message for cases involving contravariance.
* `VarianceDiagInfo` currently only handles mutable pointers/references. However, user-defined types (structs, enums, and unions) have the variance of their type parameters inferred, so it would be good to eventually display information about that. We'll want to try to find a balance between displaying too much and too little information about how the variance was inferred.
* The improved error messages are only displayed when `#![feature(nll)]` / `-Z borrowck=mir` is enabled. If issue https://github.com/rust-lang/rust/issues/58781 is not resolved relatively soon, then we might want to duplicate some of this logic in the 'current' (non-NLL) region/outlives handling code.
linker: Reorder linker arguments
- Split arguments into order-independent and order-dependent, to define more precisely what (pre-,post-,late-,)link-args mean.
- Add some comments.
- Combine all native library arguments together, to simplify potential support for library deduplication and similar things
- Split arguments into order-independent and order-dependent, to define more precisely what (pre,post,late)-link-args mean
parser: Ensure that all nonterminals have tokens after parsing
`parse_nonterminal` should always result in something with tokens.
This requirement wasn't satisfied in two cases:
- `stmt` nonterminal with expression statements (e.g. `0`, or `{}`, or `path + 1`) because `fn parse_stmt_without_recovery` forgot to propagate `force_collect` in some cases.
- `expr` nonterminal with expressions with built-in attributes (e.g. `#[allow(warnings)] 0`) due to an incorrect optimization in `fn parse_expr_force_collect`, it assumed that all expressions starting with `#` have their tokens collected during parsing, but that's not true if all the attributes on that expression are built-in and inert.
(Discovered when trying to implement eager `cfg` expansion for all attributes https://github.com/rust-lang/rust/pull/83824#issuecomment-817317170.)
r? `@Aaron1011`
Drop an `if let` that will always succeed
We've already checked that `proj_base == []` in the line above and renaming
`place_local` to `local` doesn't gain us anything.
``@rustbot`` modify labels +C-cleanup +T-compiler
Tweak wasm_base target spec to indicate linker is not GNU and update linker inferring logic for wasm-ld.
Reported via [Zulip](https://rust-lang.zulipchat.com/#narrow/stream/131828-t-compiler/topic/wasi.20linker.20unknown.20argument.3A.20--as-needed): we try passing `--as-needed` to the linker if it's GNU ld which `wasm-ld` is not. Usually this isn't an issue for wasm as we would use the WasmLd linker driver but because the linker in question (`wasm32-unknown-wasi-wasm-ld`) ended with `-ld` our linker inferring [logic](f64503eb55/compiler/rustc_codegen_ssa/src/back/link.rs (L957-L1040)) used the `GccLinker` implementations. (UPD: The linker inferring logic actually didn't apply in this case because the linker is actually invoked through gcc in the reported issue. But it's still worth updating the logic I think.)
This change then has 2 parts:
1. Update wasm_base target spec to indicate `linker_is_gnu: false` plus a few additions of `target.is_like_wasm` to handle flags `wasm-ld` does in fact support.
2. Improve the linker detection logic to properly determine the correct flavor of wasm linker we're using when we can.
We need to add the new `target.is_like_wasm` branches to handle the case where the "linker" used could be something like clang which would then under the hood call wasm-ld.
Preserve metadata w/ Solaris-like linkers.
#84468 moved the `-zignore` linker flag from the `gc_sections` method to `add_as_needed` which is more accurate but Solaris-style linkers will also end up removing an unreferenced ELF sections [1]. This had the unfortunate side effect of causing the `.rustc` section (which has the metada) to be removed which could cause issues when trying to link against the resulting crates or use proc macros.
Since the `-zignore` is positional, we fix this by moving the metadata objects to before the flag.
[1] Specifically a section is considered unreferenced if:
* The section is allocatable
* No other sections bind to (relocate) to this section
* The section provides no global symbols
https://docs.oracle.com/cd/E19683-01/817-3677/6mj8mbtbs/index.html#chapter4-19
Show test type during prints
Test output can sometimes be confusing. For example doctest with the no_run argument are displayed the same way than test that are run.
During #83857 I got the feedback that test output can be confusing.
For the moment test output is
```
test $DIR/test-type.rs - f (line 12) ... ignored
test $DIR/test-type.rs - f (line 15) ... ok
test $DIR/test-type.rs - f (line 21) ... ok
test $DIR/test-type.rs - f (line 6) ... ok
```
I propose to change output by indicating the test type as
```
test $DIR/test-type.rs - f (line 12) ... ignored
test $DIR/test-type.rs - f (line 15) - compile ... ok
test $DIR/test-type.rs - f (line 21) - compile fail ... ok
test $DIR/test-type.rs - f (line 6) ... ok
```
by indicating the test type after the test name (and in the case of doctest after the function name and line) and before the "...".
------------
Note: this is a proof of concept, the implementation is probably not optimal as the properties added in `TestDesc` are only use in the display and does not represent actual change of behavior, maybe `TestType::DocTest` could have fields
Partial support for raw-dylib linkage
First cut of functionality for issue #58713: add support for `#[link(kind = "raw-dylib")]` on `extern` blocks in lib crates compiled to .rlib files. Does not yet support `#[link_name]` attributes on functions, or the `#[link_ordinal]` attribute, or `#[link(kind = "raw-dylib")]` on `extern` blocks in bin crates; I intend to publish subsequent PRs to fill those gaps. It's also not yet clear whether this works for functions in `extern "stdcall"` blocks; I also intend to investigate that shortly and make any necessary changes as a follow-on PR.
This implementation calls out to an LLVM function to construct the actual `.idata` sections as temporary `.lib` files on disk and then links those into the generated .rlib.
BPF target support
This adds `bpfel-unknown-none` and `bpfeb-unknown-none`, two new no_std targets that generate little and big endian BPF. The approach taken is very similar to the cuda target, where `TargetOptions::obj_is_bitcode` is enabled and code generation is done by the linker.
I added the targets to `dist-various-2`. There are [some tests](https://github.com/alessandrod/bpf-linker/tree/main/tests/assembly) in bpf-linker and I'm planning to add more. Those are currently not ran as part of rust CI.
Remove special handling of `box_free` from `LocalAnalyzer`
The special casing of `box_free` predates the use of dominators in
analyzer. It is no longer necessary now that analyzer verifies that
the first assignment dominates all uses.
This addresses a codegen-issue that needs to be fixed upstream in LLVM.
While we wait for the fix, we can disable it.
Verified manually that the outliner is no longer run when
`-Copt-level=z` is specified, and also that you can override this with
`-Cllvm-args=-enable-machine-outliner` if you need it anyway.
A regression test is not really feasible in this instance, given that we
do not have any minimal reproducers.
Fixes#85351
Allow raw pointers in SIMD types
Closes#85915 by loosening the strictness in typechecking and adding a test to guarantee it passes.
This still might be too strict, as references currently do pass monomorphization, but my understanding is that they are not guaranteed to be "scalar" in the same way.
Remove `doc(include)`
This nightly feature is redundant now that `extended_key_value_attributes` is stable (https://github.com/rust-lang/rust/pull/83366). `@rust-lang/rustdoc` not sure if you think this needs FCP; there was already an FCP in #82539, but technically it was for deprecating, not removing the feature altogether.
This should not be merged before #83366.
cc `@petrochenkov`
Implement DepTrackingHash for `Option` through blanket impls instead of macros
This avoids having to add a new macro call for both the `Option` and the type itself.
Noticed this while working on https://github.com/rust-lang/rust/pull/84233.
r? `@Aaron1011`
This does not yet support #[link_name] attributes on functions, the #[link_ordinal]
attribute, #[link(kind = "raw-dylib")] on extern blocks in bin crates, or
stdcall functions on 32-bit x86.
don't suggest unsized indirection in where-clauses
Skip where-clauses when suggesting using indirection in combination with
`?Sized` bounds on type parameters.
Fixes#85943.
`@estebank` I think this doesn't conflict with your work in #85947; please let me know if you'd like me to cherry pick it to a new branch based on yours instead.
wasm: Make simd types passed via indirection again
This commit updates wasm target specs to use `simd_types_indirect: true`
again. Long ago this was added since wasm simd types were always
translated to `v128` under-the-hood in LLVM, meaning that it didn't
matter whether that target feature was enabled or not. Now, however,
`v128` is conditionally used in codegen depending on target features
enabled, meaning that it's possible to get linker errors about different
signatures in code that correctly uses simd types. The fix is the same
as for all other platforms, which is to pass the type indirectly.
rustc: Store metadata-in-rlibs in object files
This commit updates how rustc compiler metadata is stored in rlibs.
Previously metadata was stored as a raw file that has the same format as
`--emit metadata`. After this commit, however, the metadata is encoded
into a small object file which has one section which is the contents of
the metadata.
The motivation for this commit is to fix a common case where #83730
arises. The problem is that when rustc crates a `dylib` crate type it
needs to include entire rlib files into the dylib, so it passes
`--whole-archive` (or the equivalent) to the linker. The problem with
this, though, is that the linker will attempt to read all files in the
archive. If the metadata file were left as-is (today) then the linker
would generate an error saying it can't read the file. The previous
solution was to alter the rlib just before linking, creating a new
archive in a temporary directory which has the metadata file removed.
This problem from before this commit is now removed if the metadata file
is stored in an object file that the linker can read. The only caveat we
have to take care of is to ensure that the linker never actually
includes the contents of the object file into the final output. We apply
similar tricks as the `.llvmbc` bytecode sections to do this.
This involved changing the metadata loading code a bit, namely updating
some of the LLVM C APIs used to use non-deprecated ones and fiddling
with the lifetimes a bit to get everything to work out. Otherwise though
this isn't intended to be a functional change really, only that metadata
is stored differently in archives now.
This should end up fixing #83730 because by default dylibs will no
longer have their rlib dependencies "altered" meaning that
split-debuginfo will continue to have valid paths pointing at the
original rlibs. (note that we still "alter" rlibs if LTO is enabled to
remove Rust object files and we also "alter" for the #[link(cfg)]
feature, but that's rarely used).
Closes#83730
Since PR #69251, the `#[track_caller]` attribute has been supported on
traits. However, it only has an effect on direct (monomorphized) method
calls. Calling a `#[track_caller]` method on a trait object will *not*
propagate caller location information - instead, `Location::caller()` will
return the location of the method definition.
This PR forwards caller location information when `#[track_caller]` is
present on the method definition in the trait. This is possible because
`#[track_caller]` in this position is 'inherited' by any impls of that
trait, so all implementations will have the same ABI.
This PR does *not* change the behavior in the case where
`#[track_caller]` is present only on the impl of a trait.
While all implementations of the method might have an explicit
`#[track_caller]`, we cannot know this at codegen time, since other
crates may have impls of the trait. Therefore, we keep the current
behavior of not forwarding the caller location, ensuring that all
implementations of the trait will have the correct ABI.
See the modified test for examples of how this works
This commit updates how rustc compiler metadata is stored in rlibs.
Previously metadata was stored as a raw file that has the same format as
`--emit metadata`. After this commit, however, the metadata is encoded
into a small object file which has one section which is the contents of
the metadata.
The motivation for this commit is to fix a common case where #83730
arises. The problem is that when rustc crates a `dylib` crate type it
needs to include entire rlib files into the dylib, so it passes
`--whole-archive` (or the equivalent) to the linker. The problem with
this, though, is that the linker will attempt to read all files in the
archive. If the metadata file were left as-is (today) then the linker
would generate an error saying it can't read the file. The previous
solution was to alter the rlib just before linking, creating a new
archive in a temporary directory which has the metadata file removed.
This problem from before this commit is now removed if the metadata file
is stored in an object file that the linker can read. The only caveat we
have to take care of is to ensure that the linker never actually
includes the contents of the object file into the final output. We apply
similar tricks as the `.llvmbc` bytecode sections to do this.
This involved changing the metadata loading code a bit, namely updating
some of the LLVM C APIs used to use non-deprecated ones and fiddling
with the lifetimes a bit to get everything to work out. Otherwise though
this isn't intended to be a functional change really, only that metadata
is stored differently in archives now.
This should end up fixing #83730 because by default dylibs will no
longer have their rlib dependencies "altered" meaning that
split-debuginfo will continue to have valid paths pointing at the
original rlibs. (note that we still "alter" rlibs if LTO is enabled to
remove Rust object files and we also "alter" for the #[link(cfg)]
feature, but that's rarely used).
Closes#83730
Support for force-warns
Implements https://github.com/rust-lang/rust/issues/85512.
This PR adds a new command line option `force-warns` which will force the provided lints to warn even if they are allowed by some other mechanism such as `#![allow(warnings)]`.
Some remaining issues:
* https://github.com/rust-lang/rust/issues/85512 mentions that `force-warns` should also be capable of taking lint groups instead of individual lints. This is not implemented.
* If a lint has a higher warning level than `warn`, this will cause that lint to warn instead. We probably want to allow the lint to error if it is set to a higher lint and is not allowed somewhere else.
* One test is currently ignored because it's not working - when a deny-by-default lint is allowed, it does not currently warn under `force-warns`. I'm not sure why, but I wanted to get this in before the weekend.
r? `@nikomatsakis`
For extern providers, both provide and provide_extern are called.
wasm_import_module_map is already provided in provide, so it doesn't
need to be provided in provide_extern.
Reland - Report coverage `0` of dead blocks
Fixes: #84018
With `-Z instrument-coverage`, coverage reporting of dead blocks
(for example, blocks dropped because a conditional branch is dropped,
based on const evaluation) is now supported.
Note, this PR relands an earlier, reverted PR that failed when compiling
generators. The prior issues with generators has been resolved and a new
test was added to prevent future regressions.
Check out the resulting changes to test coverage of dead blocks in the
test coverage reports in this PR.
r? `@tmandry`
fyi: `@wesleywiser`
Restoring the `num_def_ids` function in the CStore API
## The context
I am the maintainer of https://github.com/hacspec/hacspec, an embedded Rust DSL aimed at cryptographic specifications. As it is normal for an embedded DSL, Hacspec's compiler relies on being plugged to the internal API of the Rust compiler, which is unstable and subject to changes.
## The problem
The Hacspec compiler features its own typechecker, that performs an additional, more restrictive typechecking pass over the Rust code of a crate. To complete this typechecking, the Hacspec compiler needs to retrieve the signature of functions defined in non-local imported crates. Rather than retrieving these signatures on-demand, the Hacspec compiler pre-populates its typechecking context with all the Hacspec-compatible symbols defined in non-local crates first. This requires having a way to iterate over all the definitions in a non-local crate.
I used to do this with `CrateMetadata::all_def_path_hashes_and_def_ids`, but this function was deleted in 908bf5a310. Then, I fellback on `CStore::num_def_ids`, exploiting the fact that all the `DefIds` for a crate have the same `krate_num` and range from `0` to `num_def_ids(krate_num)`. But `num_def_ids` was deleted in b6120bfb35.
I looked to the `Cstore::item_children_untracked` function to replicate the feature of traversing through all the `DefId` for a crate, using `CRATE_DEF_INDEX` as the root, but this does not work as recursive `Cstore::item_children_untracked` calls do not reach all the symbols I was able to reach using the two previous methods.
## Description of this PR
This PR simply restores in the public API of `CStore` the `num_def_ids` function, giving the size of the definition table for a given crate.
Remove unused feature gates
The first commit removes a usage of a feature gate, but I don't expect it to be controversial as the feature gate was only used to workaround a limitation of rust in the past. (closures never being `Clone`)
The second commit uses `#[allow_internal_unstable]` to avoid leaking the `trusted_step` feature gate usage from inside the index newtype macro. It didn't work for the `min_specialization` feature gate though.
The third commit removes (almost) all feature gates from the compiler that weren't used anyway.
This commit updates wasm target specs to use `simd_types_indirect: true`
again. Long ago this was added since wasm simd types were always
translated to `v128` under-the-hood in LLVM, meaning that it didn't
matter whether that target feature was enabled or not. Now, however,
`v128` is conditionally used in codegen depending on target features
enabled, meaning that it's possible to get linker errors about different
signatures in code that correctly uses simd types. The fix is the same
as for all other platforms, which is to pass the type indirectly.
Improve debugging experience for enums on windows-msvc
This PR makes significant improvements over the status quo of debugging enums on the windows-msvc platform with either WinDbg or Visual Studio in three ways:
1. Improves the debugger experience for directly tagged enums.
2. Fixes a bug which caused the debugger to sometimes show the wrong debug info for niche layout enums. For example, `Option<&u32>` could sometimes use the debug info for `Option<&f64>` instead leading to nonsensical variable values in the debugger.
3. Significantly improves the debugger experience for niche-layout enums.
Let's look at a few examples:
```rust
pub enum CStyleEnum {
Base = 2,
Exponent = 16,
}
pub enum NicheLayoutEnum {
Tag1,
Data { my_data: CStyleEnum },
Tag2,
Tag3,
Tag4,
}
pub enum OtherEnum<T> {
Case1(T),
Case2(T),
}
fn main() {
let a = Some(CStyleEnum::Base);
let b = Option::<CStyleEnum>::None;
let c = NicheLayoutEnum::Tag1;
let d = NicheLayoutEnum::Data { my_data: CStyleEnum::Exponent };
let e = NicheLayoutEnum::Tag2;
let f = Some(&1u32);
let g = Option::<&'static u32>::None;
let h = Some(&2u64);
let i = Option::<&'static u64>::None;
let j = Some(12u32);
let k = Option::<u32>::None;
let l = Some(12.34f64);
let m = Option::<f64>::None;
let n = CStyleEnum::Base;
let o = CStyleEnum::Exponent;
let p = Some("IAMA optional string!".to_string());
let q = OtherEnum::Case1(42u32);
}
```
This is what WinDbg Preview shows using the latest rustc nightly:
Most of the variables don't show a meaningful value expect for a few cases that we have targeted natvis definitions covering. Even worse, drilling into many of these variables shows information that can be difficult to interpret without an understanding of the layout of Rust types:
With the changes in this PR, we're able to write two natvis definitions that cover all enum cases generally. After building with these changes, WinDbg now shows this instead:
Drilling into the same variables, we can see much more useful information:
Fixes#84670Fixes#84671
Restored underlying num_def_ids_method
Update compiler/rustc_metadata/src/rmeta/decoder/cstore_impl.rs
Changed name to fit with naming convention
Co-authored-by: bjorn3 <bjorn3@users.noreply.github.com>
Update compiler/rustc_metadata/src/rmeta/decoder/cstore_impl.rs
Replace regular doc with Rustdoc comment
Co-authored-by: Joshua Nelson <jyn514@gmail.com>
Clarifies third-party use of num_def_ids_untracked
Use pattern matching instead of checking lengths explicitly
This piece of code checks that there are exaclty two variants, one having
exactly one field, the other having exactly zero fields. If any of these
conditions is violated, it returns `None`. Otherwise it assigns that one
field's ty to `field_ty`.
Instead of fiddling with indices and length checks explicitly, use pattern
matching to simplify this.
`@rustbot` modify labels +C-cleanup +T-compiler
Turn off frame pointer elimination on all Apple platforms.
This ends up disabling frame pointer elimination on aarch64_apple_darwin
which matches what clang does by default along with the
aarch64_apple_ios and x86_64_apple_darwin targets.
Further, the Apple docs "Writing ARM64 Code for Apple Platforms" has a section
called "Respect the Purpose of Specific CPU Registers" which
specifically calls out the frame pointer register (x29):
The frame pointer register (x29) must always address a valid frame
record. Some functions — such as leaf functions or tail calls — may
opt not to create an entry in this list As a result, stack traces
are always meaningful, even without debug information.
Other platforms are updated to not override the default.
rustc: Allow safe #[target_feature] on wasm
This commit updates the compiler's handling of the `#[target_feature]`
attribute when applied to functions on WebAssembly-based targets. The
compiler in general requires that any functions with `#[target_feature]`
are marked as `unsafe` as well, but this commit relaxes the restriction
for WebAssembly targets where the attribute can be applied to safe
functions as well.
The reason this is done is that the motivation for this feature of the
compiler is not applicable for WebAssembly targets. In general the
`#[target_feature]` attribute is used to enhance target CPU features
enabled beyond the basic level for the rest of the compilation. If done
improperly this means that your program could execute an instruction
that the CPU you happen to be running on does not understand. This is
considered undefined behavior where it is unknown what will happen (e.g.
it's not a deterministic `SIGILL`).
For WebAssembly, however, the target is different. It is not possible
for a running WebAssembly program to execute an instruction that the
engine does not understand. If this were the case then the program would
not have validated in the first place and would not run at all. Even if
this were allowed in some hypothetical future where engines have some
form of runtime feature detection (which they do not right now) any
implementation of such a feature would generate a trap if a module
attempts to execute an instruction the module does not understand. This
deterministic trap behavior would still not fall into the category of
undefined behavior because the trap is deterministic.
For these reasons the `#[target_feature]` attribute is now allowed on
safe functions, but only for WebAssembly targets. This notably enables
the wasm-SIMD intrinsics proposed for stabilization in #74372 to be
marked as safe generally instead of today where they're all `unsafe` due
to the historical implementation of `#[target_feature]` in the compiler.
This ends up disabling frame pointer elimination on aarch64_apple_darwin
which matches what clang does by default along with the
aarch64_apple_ios and x86_64_apple_darwin targets.
Further, the Apple docs "Writing ARM64 Code for Apple Platforms" has a section
called "Respect the Purpose of Specific CPU Registers" which
specifically calls out the frame pointer register (x29):
The frame pointer register (x29) must always address a valid frame
record. Some functions — such as leaf functions or tail calls — may
opt not to create an entry in this list As a result, stack traces
are always meaningful, even without debug information.
Other platforms are updated to not override the default.
Previously, we would generate a single struct with the layout of the
dataful variant plus an extra field whose name contained the value of
the niche (this would only really work for things like `Option<&_>`
where we can determine that the `None` case maps to `0` but for enums
that have multiple tag only variants, this doesn't work).
Now, we generate a union of two structs, one which is the layout of the
dataful variant and one which just has a way of reading the
discriminant. We also generate an enum which maps the discriminant value
to the tag only variants.
We also encode information about the range of values which correspond to
the dataful variant in the type name and then use natvis to determine
which union field we should display to the user.
As a result of this change, all niche-layout enums render correctly in
WinDbg and Visual Studio!
This wasn't necessary for msvc and caused issues where different types
with the same name such as different instantiations of `Option<T>` would
have colliding debuginfo. This confused the debugger which would pick
one of the type definitions and use for all types with that name even
though they had different layout.
Avoid creating anonymous nodes with zero or one dependency.
Anonymous nodes are only useful to encode dependencies, and cannot be replayed from one compilation session to another.
As such, anonymous nodes without dependency are always green.
Anonymous nodes with only one dependency are equivalent to this dependency.
cc #45408
cc `@michaelwoerister`
This piece of code checks that there are exaclty two variants, one having
exactly one field, the other having exactly zero fields. If any of these
conditions is violated, it returns `None`. Otherwise it assigns that one
field's ty to `field_ty`.
Instead of fiddling with indices and length checks explicitly, use pattern
matching to simplify this.
Fixes: #84018
With `-Z instrument-coverage`, coverage reporting of dead blocks
(for example, blocks dropped because a conditional branch is dropped,
based on const evaluation) is now supported.
Note, this PR relands an earlier, reverted PR that failed when compiling
generators. The prior issues with generators has been resolved and a new
test was added to prevent future regressions.
Check out the resulting changes to test coverage of dead blocks in the
test coverage reports in this PR.
Reduce the amount of untracked state in TyCtxt
Access to untracked global state may generate instances of #84970.
The GlobalCtxt contains the lowered HIR, the resolver outputs and interners.
By wrapping the resolver inside a query, we make sure those accesses are properly tracked.
As a no_hash query, all dependent queries essentially become `eval_always`,
what they should have been from the beginning.
Don't sort a `Vec` before computing its `DepTrackingHash`
Previously, we sorted the vec prior to hashing, making the hash
independent of the original (command-line argument) order. However, the
original vec was still always kept in the original order, so we were
relying on the rest of the compiler always working with it in an
'order-independent' way.
This assumption was not being upheld by the `native_libraries` query -
the order of the entires in its result depends on the order of entries
in `Options.libs`. This lead to an 'unstable fingerprint' ICE when the
`-l` arguments were re-ordered.
This PR removes the sorting logic entirely. Re-ordering command-line
arguments (without adding/removing/changing any arguments) seems like a
really niche use case, and correctly optimizing for it would require
additional work. By always hashing arguments in their original order, we
can entirely avoid a cause of 'unstable fingerprint' errors.
Emit a hard error when a panic occurs during const-eval
Previous, a panic during const evaluation would go through the
`const_err` lint. This PR ensures that such a panic always causes
compilation to fail.
Fix span of redundant generic arguments
Fixes#71563
Above issue is about lifetime arguments, but generic arguments also have same problem.
This PR fixes both help messages.
The special casing of `box_free` predates the use of dominators in
analyzer. It is no longer necessary now that analyzer verifies that
the first assignment dominates all uses.
Use correct edition when parsing `:pat` matchers
As described in issue #85708, we currently do not properly decode
`SyntaxContext::root()` and `ExpnId::root()` from foreign crates. As a
result, when we decode a span from a foreign crate with
`SyntaxContext::root()`, we end up up considering it to have the edition
of the *current* crate, instead of the foreign crate where it was
originally created.
A full fix for this issue will be a fairly significant undertaking.
Fortunately, it's possible to implement a partial fix, which gives us
the correct edition-dependent behavior for `:pat` matchers when the
macro is loaded from another crate. Since we have the edition of the
macro's defining crate available, we can 'recover' from seeing a
`SyntaxContext::root()` and use the edition of the macro's defining
crate.
Any solution to issue #85708 must reproduce the behavior of this
targeted fix - properly preserving a foreign `SyntaxContext::root()`
means (among other things) preserving its edition, which by definition
is the edition of the foreign crate itself. Therefore, this fix moves us
closer to the correct overall solution, and does not expose any new
incorrect behavior to macros.
Use command line metadata path if provided
If the command-line has `--emit metadata=some/path/libfoo.rmeta` then
use that.
Closes#85356
I couldn't find any existing tests for the `--emit TYPE=PATH` command line syntax, so I wasn't sure how to test this aside from ad-hoc manual testing. Is there a ui test type for "generated output file with expected name"?
Merge CrateDisambiguator into StableCrateId
This simplifies the code and potentially improves performance by reducing the amount of hashed data.
Fixes https://github.com/rust-lang/rust/issues/85795
Make `Step` trait safe to implement
This PR makes a few modifications to the `Step` trait that I believe better position it for stabilization in the short term. In particular,
1. `unsafe trait TrustedStep` is introduced, indicating that the implementation of `Step` for a given type upholds all stated invariants (which have remained unchanged). This is gated behind a new `trusted_step` feature, as stabilization is realistically blocked on min_specialization.
2. The `Step` trait is internally specialized on the `TrustedStep` trait, which avoids a serious performance regression.
3. `TrustedLen` is implemented for `T: TrustedStep` as the latter's invariants subsume the former's.
4. The `Step` trait is no longer `unsafe`, as the invariants must not be relied upon by unsafe code (unless the type implements `TrustedStep`).
5. `TrustedStep` is implemented for all types that implement `Step` in the standard library and compiler.
6. The `step_trait_ext` feature is merged into the `step_trait` feature. I was unable to find any reasoning for the features being split; the `_unchecked` methods need not necessarily be stabilized at the same time, but I think it is useful to have them under the same feature flag.
All existing implementations of `Step` will be broken, as it is not possible to `unsafe impl` a safe trait. Given this trait only exists on nightly, I feel this breakage is acceptable. The blanket `impl<T: Step> TrustedLen for T` will likely cause some minor breakage, but this should be covered by the equivalent impl for `TrustedStep`.
Hopefully these changes are sufficient to place `Step` in decent position for stabilization, which would allow user-defined types to be used with `a..b` syntax.
Don't panic when failing to initialize incremental directory.
This removes a panic when rustc fails to initialize the incremental directory. This can commonly happen on various filesystems that don't support locking (often various network filesystems). Panics can be confusing and scary, and there are already plenty of issues reporting this.
This has been panicking since 1.22 due to I think #44502 which was a major rework of how things work. Previously, things were simpler and the [`load_dep_graph`](https://github.com/rust-lang/rust/blob/1.21.0/src/librustc_incremental/persist/load.rs#L43-L65) function would emit an error and then continue on without panicking. With 1.22, [`load_dep_graph`](https://github.com/rust-lang/rust/blob/1.22.0/src/librustc_incremental/persist/load.rs#L44) was changed so that it assumes it can load the data without errors. Today, the problem is that it calls [`prepare_session_directory`](fbf1b1a719/compiler/rustc_interface/src/passes.rs (L175-L179)) and then immediately calls `garbage_collect_session_directories` which will panic since the session is `IncrCompSession::NotInitialized`.
The solution here is to have `prepare_session_directory` return an error that must be handled so that compilation stops if it fails.
Some other options:
* Ignore directory lock failures.
* Print a warning on directory lock failure, but otherwise continue with incremental enabled.
* Print a warning on directory lock failure, and disable incremental.
* Provide a different locking mechanism.
Cargo ignores lock errors if locking is not supported, so that would be a precedent for the first option. These options would require quite a bit more changes, but I'm happy to entertain any of them, as I think they all have valid justifications.
There is more discussion on the many issues where this is reported: #49773, #59224, #66513, #76251. I'm not sure if this can be considered closing any of those, though, since I think there is some value in discussing if there is a way to avoid the error altogether. But I think it would make sense to at least close all but one to consolidate them.
As described in issue #85708, we currently do not properly decode
`SyntaxContext::root()` and `ExpnId::root()` from foreign crates. As a
result, when we decode a span from a foreign crate with
`SyntaxContext::root()`, we end up up considering it to have the edition
of the *current* crate, instead of the foreign crate where it was
originally created.
A full fix for this issue will be a fairly significant undertaking.
Fortunately, it's possible to implement a partial fix, which gives us
the correct edition-dependent behavior for `:pat` matchers when the
macro is loaded from another crate. Since we have the edition of the
macro's defining crate available, we can 'recover' from seeing a
`SyntaxContext::root()` and use the edition of the macro's defining
crate.
Any solution to issue #85708 must reproduce the behavior of this
targeted fix - properly preserving a foreign `SyntaxContext::root()`
means (among other things) preserving its edition, which by definition
is the edition of the foreign crate itself. Therefore, this fix moves us
closer to the correct overall solution, and does not expose any new
incorrect behavior to macros.
A bit more polish on const eval errors
This PR adds a bit more polish to the const eval errors:
- a slight improvement to the PME messages from #85633: I mentioned there that the erroneous item's paths were dependent on the environment, and could be displayed fully qualified or not. This can obscure the items when they come from a dependency. This PR uses the pretty-printing code ensuring the items' paths are not trimmed.
- whenever there are generics involved in an item where const evaluation errors out, the error message now displays the instance and its const arguments, so that we can see which instantiated item and compile-time values lead to the error.
So we get this slight improvement for our beloved `stdarch` example, on nightly:
```
error[E0080]: evaluation of constant value failed
--> ./stdarch/crates/core_arch/src/macros.rs:8:9
|
8 | assert!(IMM >= MIN && IMM <= MAX, "IMM value not in expected range");
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the evaluated program panicked at 'IMM value not in expected range', /rustc/9111b8ae9793f18179a1336417618fc07a9cac85/library/core/src/../../stdarch/crates/core_arch/src/macros.rs:8:9
|
```
to this PR's:
```
error[E0080]: evaluation of `core::core_arch::macros::ValidateConstImm::<51_i32, 0_i32, 15_i32>::VALID` failed
--> ./stdarch/crates/core_arch/src/macros.rs:8:9
|
8 | assert!(IMM >= MIN && IMM <= MAX, "IMM value not in expected range");
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the evaluated program panicked at 'IMM value not in expected range', ./stdarch/crates/core_arch/src/macros.rs:8:9
|
```
with this PR.
Of course this is an idea from Oli, so maybe r? `@oli-obk` if they have the time.
This commit updates the compiler's handling of the `#[target_feature]`
attribute when applied to functions on WebAssembly-based targets. The
compiler in general requires that any functions with `#[target_feature]`
are marked as `unsafe` as well, but this commit relaxes the restriction
for WebAssembly targets where the attribute can be applied to safe
functions as well.
The reason this is done is that the motivation for this feature of the
compiler is not applicable for WebAssembly targets. In general the
`#[target_feature]` attribute is used to enhance target CPU features
enabled beyond the basic level for the rest of the compilation. If done
improperly this means that your program could execute an instruction
that the CPU you happen to be running on does not understand. This is
considered undefined behavior where it is unknown what will happen (e.g.
it's not a deterministic `SIGILL`).
For WebAssembly, however, the target is different. It is not possible
for a running WebAssembly program to execute an instruction that the
engine does not understand. If this were the case then the program would
not have validated in the first place and would not run at all. Even if
this were allowed in some hypothetical future where engines have some
form of runtime feature detection (which they do not right now) any
implementation of such a feature would generate a trap if a module
attempts to execute an instruction the module does not understand. This
deterministic trap behavior would still not fall into the category of
undefined behavior because the trap is deterministic.
For these reasons the `#[target_feature]` attribute is now allowed on
safe functions, but only for WebAssembly targets. This notably enables
the wasm-SIMD intrinsics proposed for stabilization in #74372 to be
marked as safe generally instead of today where they're all `unsafe` due
to the historical implementation of `#[target_feature]` in the compiler.
const-eval: disallow unwinding across functions that `!fn_can_unwind()`
Following https://github.com/rust-lang/miri/pull/1776#discussion_r633074343, so r? `@RalfJung`
This PR turns `unwind` in `StackPopCleanup::Goto` into a new enum `StackPopUnwind`, with a `NotAllowed` variant to indicate that unwinding is not allowed. This variant is chosen based on `rustc_middle::ty::layout::fn_can_unwind()` in `eval_fn_call()` when pushing the frame. A check is added in `unwind_to_block()` to report UB if unwinding happens across a `StackPopUnwind::NotAllowed` frame.
Tested with Miri `HEAD` with [minor changes](https://github.com/rust-lang/miri/compare/HEAD..9cf3c7f0d86325a586fbcbf2acdc9232b861f1d8) and the rust-lang/miri#1776 branch with [these changes](d866c1c52f..626638fbfe).
Fix incorrect suggestions for E0605
Fixes#84598. Here is a simplified version of the problem presented in issue #84598:
```Rust
#![allow(unused_variables)]
#![allow(dead_code)]
trait T { fn t(&self) -> i32; }
unsafe fn foo(t: *mut dyn T) {
(t as &dyn T).t();
}
fn main() {}
```
The current output is:
```
error[E0605]: non-primitive cast: `*mut (dyn T + 'static)` as `&dyn T`
--> src/main.rs:7:5
|
7 | (t as &dyn T).t();
| ^^^^^^^^^^^^^ invalid cast
|
help: borrow the value for the cast to be valid
|
7 | (&t as &dyn T).t();
| ^
```
This is incorrect, though: The cast will _not_ be valid when writing `&t` instead of `t`:
```
error[E0277]: the trait bound `*mut (dyn T + 'static): T` is not satisfied
--> t4.rs:7:6
|
7 | (&t as &dyn T).t();
| ^^ the trait `T` is not implemented for `*mut (dyn T + 'static)`
|
= note: required for the cast to the object type `dyn T`
```
The correct suggestion is `&*t`, which I have implemented in this pull request. Of course, this suggestion will always require an unsafe block, but arguably, that's what the user really wants if they're trying to cast a pointer to a reference.
In any case, claiming that the cast will be valid after implementing the suggestion is overly optimistic, as the coercion logic doesn't seem to resolve all nested obligations, i.e. the cast may still be invalid after implementing the suggestion. I have therefore rephrased the suggestion slightly ("consider borrowing the value" instead of "borrow the value for the cast to be valid").
Additionally, I have fixed another incorrect suggestion not mentioned in #84598, which relates to casting immutable references to mutable ones:
```rust
fn main() {
let mut x = 0;
let m = &x as &mut i32;
}
```
currently leads to
```
error[E0605]: non-primitive cast: `&i32` as `&mut i32`
--> t5.rs:3:13
|
3 | let m = &x as &mut i32;
| ^^^^^^^^^^^^^^ invalid cast
|
help: borrow the value for the cast to be valid
|
3 | let m = &mut &x as &mut i32;
| ^^^^
```
which is obviously incorrect:
```
error[E0596]: cannot borrow data in a `&` reference as mutable
--> t5.rs:3:13
|
3 | let m = &mut &x as &mut i32;
| ^^^^^^^ cannot borrow as mutable
```
I've changed the suggestion to a note explaining the problem:
```
error[E0605]: non-primitive cast: `&i32` as `&mut i32`
--> t5.rs:3:13
|
3 | let m = &x as &mut i32;
| ^^^^^^^^^^^^^^ invalid cast
|
note: this reference is immutable
--> t5.rs:3:13
|
3 | let m = &x as &mut i32;
| ^^
note: trying to cast to a mutable reference type
--> t5.rs:3:19
|
3 | let m = &x as &mut i32;
| ^^^^^^^^
```
In this example, it would have been even nicer to suggest replacing `&x` with `&mut x`, but this would be much more complex because we would have to take apart the expression to be cast (currently, we only look at its type), and `&x` could be stored in a variable, where such a suggestion would not even be directly applicable:
```rust
fn main() {
let mut x = 0;
let r = &x;
let m = r as &mut i32;
}
```
My solution covers this case, too.
Sync rustc_codegen_cranelift
The main highlight this sync is the removal of several dependencies, making compilation of cg_clif itself faster. There have also been a couple of new features like `#[link_section]` now supporting different segments for Mach-O binaries (thanks `@eggyal!)` and the `imported_main` feature, which is currently unstable.
r? `@ghost`
`@rustbot` label +A-codegen +A-cranelift +T-compiler
Update cc
Recent commits have improved `cc`'s finding of MSVC tools on Windows. In particular it should help to address these issues: #83043 and #43468
readd capture disjoint fields gate
This readds a feature gate guard that was added in PR #83521. (Basically, there were unintended consequences to the code exposed by removing the feature gate guard.)
The root bug still remains to be resolved, as discussed in issue #85561. This is just a band-aid suitable for a beta backport.
Cc issue #85435
Note that the latter issue is unfixed until we backport this (or another fix) to 1.53 beta
stabilize member constraints
Stabilizes the use of "member constraints" in solving `impl Trait` bindings. This is a step towards stabilizing a "MVP" of "named impl Trait".
# Member constraint stabilization report
| Info | |
| --- | --- |
| Tracking issue | [rust-lang/rust#61997](https://github.com/rust-lang/rust/issues/61997) |
| Implementation history | [rust-lang/rust#61775] |
| rustc-dev-guide coverage | [link](https://rustc-dev-guide.rust-lang.org/borrow_check/region_inference/member_constraints.html) |
| Complications | [rust-lang/rust#61773] |
[rust-lang/rust#61775]: https://github.com/rust-lang/rust/pull/61775
[rust-lang/rust#61773]: https://github.com/rust-lang/rust/issues/61773
## Background
Member constraints are an extension to our region solver that was introduced to make async fn region solving tractable. There are used in situations like the following:
```rust
fn foo<'a, 'b>(...) -> impl Trait<'a, 'b> { .. }
```
The problem here is that every region R in the hidden type must be equal to *either* `'a` *or* `'b` (or `'static`). This cannot be expressed simply via 'outlives constriants' like `R: 'a`. Therefore, we introduce a 'member constraint' `R member of ['a, 'b]`.
These constraints were introduced in [rust-lang/rust#61775]. At the time, we kept them feature gated and used them only for `impl Trait` return types that are derived from `async fn`. The intention, however, was always to support them in other contexts once we had time to gain more experience with them.
**In the time since their introduction, we have encountered no surprises or bugs due to these member constraints.** They are tested extensively as part of every async function that involves multiple unrelated lifetimes in its arguments.
## Tests
The behavior of member constraints is covered by the following tests:
* [`src/test/ui/async-await/multiple-lifetimes`](20e032e650/src/test/ui/async-await/multiple-lifetimes) -- tests using the async await, which are mostly already stabilized
* [`src/test/ui/impl-trait/multiple-lifetimes.rs`](20e032e650/src/test/ui/impl-trait/multiple-lifetimes.rs)
* [`src/test/ui/impl-trait/multiple-lifetimes/ordinary-bounds-unsuited.rs`](20e032e650/src/test/ui/impl-trait/multiple-lifetimes/ordinary-bounds-unsuited.rs)
* [`src/test/ui/async-await/multiple-lifetimes/ret-impl-trait-fg.rs`](20e032e650/src/test/ui/async-await/multiple-lifetimes/ret-impl-trait-fg.rs)
* [`src/test/ui/async-await/multiple-lifetimes/ret-impl-trait-one.rs`](20e032e650/src/test/ui/async-await/multiple-lifetimes/ret-impl-trait-one.rs)
These tests cover a number of scenarios:
* `-> implTrait<'a, 'b>` with unrelated lifetimes `'a` and `'b`, as described above
* `async fn` that returns an `impl Trait` like the previous case, which desugars to a kind of "nested" impl trait like `impl Future<Output = impl Trait<'a, 'b>>`
## Potential concerns
There is a potential interaction with `impl Trait` on local variables, described in [rust-lang/rust#61773]. The challenge is that if you have a program like:
```rust=
trait Foo<'_> { }
impl Foo<'_> for &u32 { }
fn bar() {
let x: impl Foo<'_> = &44; // let's call the region variable for `'_` `'1`
}
```
then we would wind up with `'0 member of ['1, 'static]`, where `'0` is the region variable in the hidden type (`&'0 u32`) and `'1` is the region variable in the bounds `Foo<'1>`. This is tricky because both `'0` and `'1` are being inferred -- so making them equal may have other repercussions.
That said, `impl Trait` in bindings are not stable, and the implementation is pretty far from stabilization. Moreover, the difficulty highlighted here is not due to the presence of member constraints -- it's inherent to the design of the language. In other words, stabilizing member constraints does not actually cause us to accept anything that would make this problem any harder.
So I don't see this as a blocker to stabilization of member constraints; it is potentially a blocker to stablization of `impl trait` in let bindings.
E0599 suggestions and elision of generic argument if no canditate is found
fixes#81576
changes: In error E0599 (method not found) generic argument are eluded if the method was not found anywhere. If the method was found in another inherent implementation suggest that it was found elsewhere.
Example
```rust
struct Wrapper<T>(T);
struct Wrapper2<T> {
x: T,
}
impl Wrapper2<i8> {
fn method(&self) {}
}
fn main() {
let wrapper = Wrapper(i32);
wrapper.method();
let wrapper2 = Wrapper2{x: i32};
wrapper2.method();
}
```
```
Error[E0599]: no method named `method` found for struct `Wrapper<_>` in the current scope
....
error[E0599]: no method named `method` found for struct `Wrapper2<i32>` in the current scope
...
= note: The method was found for Wrapper2<i8>.
```
I am not very happy with the ```no method named `test` found for struct `Vec<_, _>` in the current scope```. I think it might be better to show only one generic argument `Vec<_>` if there is a default one. But I haven't yet found a way to do that,
While stdlib implementations of the unchecked methods require unchecked
math, there is no reason to gate it behind this for external users. The
reasoning for a separate `step_trait_ext` feature is unclear, and as
such has been merged as well.
Post-monomorphization errors traces MVP
This PR works towards better diagnostics for the errors encountered in #85155 and similar.
We can encounter post-monomorphization errors (PMEs) when collecting mono items. The current diagnostics are confusing for these cases when they happen in a dependency (but are acceptable when they happen in the local crate).
These kinds of errors will be more likely now that `stdarch` uses const generics for its intrinsics' immediate arguments, and validates these const arguments with a mechanism that triggers such PMEs.
(Not to mention that the errors happen during codegen, so only when building code that actually uses these code paths. Check builds don't trigger them, neither does unused code)
So in this PR, we detect these kinds of errors during the mono item graph walk: if any error happens while collecting a node or its neighbors, we print a diagnostic about the current collection step, so that the user has at least some context of which erroneous code and dependency triggered the error.
The diagnostics for issue #85155 now have this note showing the source of the erroneous const argument:
```
note: the above error was encountered while instantiating `fn std::arch::x86_64::_mm_blend_ps::<51_i32>`
--> issue-85155.rs:11:24
|
11 | let _blended = _mm_blend_ps(a, b, 0x33);
| ^^^^^^^^^^^^^^^^^^^^^^^^
error: aborting due to previous error
```
Note that #85155 is a reduced version of a case happening in the wild, to indirect users of the `rustfft` crate, as seen in https://github.com/ejmahler/RustFFT/issues/74. The crate had a few of these out-of-range immediates. Here's how the diagnostics in this PR would have looked on one of its examples before it was fixed:
<details>
```
error[E0080]: evaluation of constant value failed
--> ./stdarch/crates/core_arch/src/macros.rs:8:9
|
8 | assert!(IMM >= MIN && IMM <= MAX, "IMM value not in expected range");
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the evaluated program panicked at 'IMM value not in expected range', ./stdarch/crates/core_arch/src/macros.rs:8:9
|
= note: this error originates in the macro `$crate::panic::panic_2015` (in Nightly builds, run with -Z macro-backtrace for more info)
note: the above error was encountered while instantiating `fn _mm_blend_ps::<51_i32>`
--> /tmp/RustFFT/src/avx/avx_vector.rs:1314:23
|
1314 | let blended = _mm_blend_ps(rows[0], rows[2], 0x33);
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
note: the above error was encountered while instantiating `fn _mm_permute_pd::<5_i32>`
--> /tmp/RustFFT/src/avx/avx_vector.rs:1859:9
|
1859 | _mm_permute_pd(self, 0x05)
| ^^^^^^^^^^^^^^^^^^^^^^^^^^
note: the above error was encountered while instantiating `fn _mm_permute_pd::<15_i32>`
--> /tmp/RustFFT/src/avx/avx_vector.rs:1863:32
|
1863 | (_mm_movedup_pd(self), _mm_permute_pd(self, 0x0F))
| ^^^^^^^^^^^^^^^^^^^^^^^^^^
error: aborting due to previous error
For more information about this error, try `rustc --explain E0080`.
error: could not compile `rustfft`
To learn more, run the command again with --verbose.
```
</details>
I've developed and discussed this with them, so maybe r? `@oli-obk` -- but feel free to redirect to someone else of course.
(I'm not sure we can say that this PR definitely closes issue 85155, as it's still unclear exactly which diagnostics and information would be interesting to report in such cases -- and we've discussed printing backtraces before. I have prototypes of some complete and therefore noisy backtraces I showed Oli, but we decided to not include them in this PR for now)
Disallow shadowing const parameters
This pull request fixes#85348. Trying to shadow a `const` parameter as follows:
```rust
fn foo<const N: i32>() {
let N @ _ = 0;
}
```
currently causes an ICE. With my changes, I get:
```
error[E0530]: let bindings cannot shadow const parameters
--> test.rs:2:9
|
1 | fn foo<const N: i32>() {
| - the const parameter `N` is defined here
2 | let N @ _ = 0;
| ^ cannot be named the same as a const parameter
error: aborting due to previous error
```
This is the same error you get when trying to shadow a constant:
```rust
const N: i32 = 0;
let N @ _ = 0;
```
```
error[E0530]: let bindings cannot shadow constants
--> src/lib.rs:3:5
|
2 | const N: i32 = 0;
| ----------------- the constant `N` is defined here
3 | let N @ _ = 0;
| ^ cannot be named the same as a constant
error: aborting due to previous error
```
The reason for disallowing shadowing in both cases is described [here](https://github.com/rust-lang/rust/issues/33118#issuecomment-233962221) (the comment there only talks about constants, but the same reasoning applies to `const` parameters).
