I also added `// skip-codegen` to each one, to address potential concerns
that this change would otherwise slow down our test suite spending time
generating code for files that are really just meant to be checks of
compiler diagnostics.
(However, I will say: My preference is to not use `// skip-codegen` if
one can avoid it. We can use all the testing of how we drive LLVM that
we can get...)
(Updated post rebase.)
We avoid an ICE by checking for an empty meta-item list before we
index into the meta-items, and leave commentary about where we'd like
to issue unused-attributes lints in the future. Note that empty lint
attributes are already accepted by the stable compiler; generalizing
this to weird reason-only lint attributes seems like the
conservative/consilient generalization.
Vadim Petrochenkov suggested this in review ("an error? just to be
conservative"), and it turns out to be convenient from the
implementer's perspective: in the initial proposed implementation (or
`HEAD~2`, as some might prefer to call it), we were doing an entire
whole iteration over the meta items just to find the reason (before
iterating over them to set the actual lint levels). This way, we can
just peek at the end rather than adding that extra loop (or
restructuring the existing code). The RFC doesn't seem to take a
position on this, and there's some precedent for restricting things to
be at the end of a sequence (we only allow `..` at the end of a struct
pattern, even if it would be possible to let it appear anywhere in the
sequence).
This is just for the `reason =` name-value meta-item; the
`#[expect(lint)]` attribute also described in the RFC is a problem for
another day.
The place where we were directly calling `emit()` on a match block
(whose arms returned a mutable reference to a diagnostic-builder) was
admittedly cute, but no longer plausibly natural after adding the
if-let to the end of the `LintSource::Node` arm.
This regards #54503.
in which unused-parens suggestions heed what the user actually wrote
Aaron Hill pointed out that unnecessary parens around a macro call (paradigmatically, `format!`) yielded a suggestion of hideous macro-expanded code. `span_to_snippet` is fallable as far as the type system is concerned, so the pretty-printing can live on in the oft-neglected `else` branch.
Resolves#55109.
Aaron Hill pointed out that unnecessary parens around a macro call
(paradigmatically, `format!`) yielded a suggestion of hideous
macro-expanded code. (The slightly unusual choice of using the
pretty-printer to compose suggestions was quite recently commented on
in the commit message for 1081bbbfc ("abolish ICE when pretty-printing
async block"), but without any grounds to condemn it as a 𝘣𝘢𝘥
choice. Hill's report provides the grounds.) `span_to_snippet` is
fallable as far as the type system is concerned (because, who knows,
macros or something), so the pretty-printing can live on in the
oft-neglected `else` branch.
Resolves#55109.
* `ui/lifetimes/lifetime-errors/ex3-both-anon-regions-both-are-structs-4`
and `ex3-both-anon-regions-both-are-structs-3`
* `ui/lint/lint-group-style` and `lint-group-nonstandard-style`
Fix dead code lint for functions using impl Trait
Fixes https://github.com/rust-lang/rust/issues/54754
This is a minimal fix that doesn't add any new queries or touches unnecessary code. Please nominate for beta backport if wanted.
This commit updates the compiler to allow the `#[no_mangle]` (and
`#[export_name]` attributes) to be located anywhere within a crate.
These attributes are unconditionally processed, causing the compiler to
always generate an exported symbol with the appropriate name.
After some discussion on #54135 it was found that not a great reason
this hasn't been allowed already, and it seems to match the behavior
that many expect! Previously the compiler would only export a
`#[no_mangle]` symbol if it were *publicly reachable*, meaning that it
itself is `pub` and it's otherwise publicly reachable from the root of
the crate. This new definition is that `#[no_mangle]` *is always
reachable*, no matter where it is in a crate or whether it has `pub` or
not.
This should make it much easier to declare an exported symbol with a
known and unique name, even when it's an internal implementation detail
of the crate itself. Note that these symbols will persist beyond LTO as
well, always making their way to the linker.
Along the way this commit removes the `private_no_mangle_functions` lint
(also for statics) as there's no longer any need to lint these
situations. Furthermore a good number of tests were updated now that
symbol visibility has been changed.
Closes#54135
librustc_lint: In recursion warning, change 'recurring' to 'recursing'
The existing wording seems incorrect.
Aside: This warning, 'function cannot return without recursing' is not perfectly clear - it implies that the function _can_ return, it's just got to recurse. But really the fn cannot return period. Clearer wording: 'function recurses infinitely; it cannot return'; or 'function is infinitely self-recursive; it cannot return, and this is probably an error'. I like that.
Andrew Chin recently pointed out (rust-lang/cargo#5846) that it's
surprising that `cargo fix` (now shipping with Cargo itself!) doesn't
fix very common lint warnings, which is as good of a reminder as any
that we should finish #50723.
This commit polishes off this new function to compile on newer rustc as well as
update and add a suite of test cases to work with this new check for lints.
If the item is `pub`, one imagines users being confused as to why it's
not reachable/exported; a code suggestion is beyond our local knowledge
here, but we can at least offer a prose hint. (Thanks to Vadim
Petrochenkov for shooting down the present author's original bad idea
for the note text.)
While we're here, use proper HELP expectations instead of ad hoc
comments to communicate (and now, enforce) the expected suggestions in
test/ui/lint/suggestions.rs.
This is probably quite a lot less likely to come up in practice than the
"inherited" (no visibility keyword) case, but now that we have
visibility spans in the HIR, we can do this, and it presumably doesn't
hurt to be exhaustive. (Who can say but that the attention to detail
just might knock someone's socks off, someday, somewhere?)
This is inspired by #47383.
Our implementation ends up changing the `PatKind::Range` variant in the
AST to take a `Spanned<RangeEnd>` instead of just a `RangeEnd`, because
the alternative would be to try to infer the span of the range operator
from the spans of the start and end subexpressions, which is both
hideous and nontrivial to get right (whereas getting the change to the
AST right was a simple game of type tennis).
This is concerning #51043.
