Turns out that `doc_markdown` uses a non-cheap rustdoc function to
convert from markdown ranges into source spans. And it was using it a
lot (about once every 17 lines of documentation on `tokio`, which ends
up being about 2000 times).
This ended up being about 18% of the total Clippy runtime as discovered
by lintcheck --perf in docs-heavy crates. This PR optimizes one of the
cases in which Clippy calls the function, and a future PR once
pulldown-cmark/pulldown-cmark#1034 is merged will be opened. This PR
lands the use of the function into the single-digit zone.
Note that not all crates were affected by this crate equally, those with
more docs are affected far more than those light ones.
changelog:[`clippy::doc_markdown`] has been optimized by 50%
This stops using `cargo fmt` and instead calls rustfmt directly with the
list of all files.
All `cargo fmt` does is find the crate roots and passes the edition from
`cargo.toml`. Since the edition is set in `rustfmt.toml` for the test
files and we're already iterating through all the files this is not
needed.
`--skip-children` is used since we already pass all the files, so the
automatic detection isn't buying us anything other than running slower.
~Second commit~ (part of the first commit now) is a change to only use
the `ignore` option in `rustfmt.toml` rather than having a way in `cargo
dev fmt` to ignore files.
r? @samueltardieu
changelog: none
Fixes
https://rust-lang.zulipchat.com/#narrow/stream/257328-clippy/topic/Ambiguous.20default.20value.20for.20.60trivial-copy-size-limit.60
The current situation is
| Target width | `trivial-copy-size-limit`|
|--------|--------|
| 8-bit | 2 |
| 16-bit | 4 |
| 32-bit | 8 |
| 64-bit | 8 |
~~Since practically speaking it's almost always 8, let's go with that as
the unconditional default to make it easier to understand~~
Now defaults to `target_pointer_width`
changelog: [`trivial-copy-size-limit`] now also defaults to the size of
a target pointer (unchanged for 64-bit targets)
The first commit fixesrust-lang/rust-clippy#14802: when a slice is
directly present, it must be dereferenced (instead of referenced -1
times) before being passed to `mem::size_of_val()`.
The second commit triggers the lint in a `const` contact when MSRV ≥
1.85.
changelog: [`manual_slice_size_computation`]: fix ICE in suggestion to
efficiently compute the size of a slice, and trigger the lint in `const`
context as well
The `explicit_into_iter_loop`, `explicit_iter_loop` and `iter_next_loop`
will now:
- trigger only when the triggering expression is not located into macro
code;
- properly expose code rewrite proposal with code coming from the root
context.
changelog: [`explicit_into_iter_loop`, `explicit_iter_loop`,
`iter_next_loop`]: behave in macro context
Fixesrust-lang/rust-clippy#14630
Fixesrust-lang/rust-clippy#14750 by reducing the lint span for
needless_return.
changelog: [`needless_return`]: Lint span no longer wraps to previous
line
If a type is incomplete, for example if generic parameters are not
available yet, although they are not escaping, its layout may not be
computable. Calling `TyCtxt::layout_of()` would create a delayed bug in
the compiler.
changelog: [`zero_sized_map_values`]: fix ICE
Fixesrust-lang/rust-clippy#14822
r? @Jarcho
If a type is incomplete, for example if generic parameters are not
available yet, although they are not escaping, its layout may not
be computable. Calling `TyCtxt::layout_of()` would create a delayed bug
in the compiler.
The `explicit_into_iter_loop`, `explicit_iter_loop` and `iter_next_loop`
will now:
- trigger only when the triggering expression is not located into macro
code;
- properly expose code rewrite proposal with code coming from the root
context.
A `return` in an expression makes it divergent and cannot be removed
blindly. While this stripping might have been introduced as a way to
catch more cases, it was improperly used, and no tests exhibit a failure
when this special handling is removed.
changelog: [`needless_match`]: do not strip `return` as it might make
the `if let` or `match` divergent in some cases
Fixesrust-lang/rust-clippy#14754
This patch series makes heavy use of interned symbols when matching
against known method names:
- the first commit reorders the current list of symbols in
`clippy_utils::sym`
- the second commit adds symbol reordering and order checking to `clippy
dev fmt` / `clippy dev fmt --check`
- the third commit converts many uses of string matching during linting
to symbols matching
The symbols are kept as-is (not rendered as strings) as much as possible
to avoid needing locking the interner as much as possible. Static
strings have been kept when they are only used when emitting a
diagnostic, as there is no benefit in using interned strings for
de-interning them right after.
changelog: none
r? @Alexendoo
Items such as the `declare_tool_lint!()` macro are publicly reexported
from `rustc_lint_defs` to `rustc_session`. Use the latter for
consistency.
changelog: none
Fixes#13793.
Interestingly enough, to actually check that the macro call has at least
two arguments, we need to use the rust lexer after getting the original
source code snippet.
changelog: Add new `useless_concat` lint
It is not enough to check if an expression type implements `Drop` to
determine whether it can have a significant side-effect.
Also, add tests for unchecked cases which were explicitly handled in the
code, such as checking for side effect in a `struct`'s fields, or in its
base expression.
Fixrust-lang/rust-clippy#14592
changelog: [`no_effect_underscore_binding`]: do not propose to remove
the assignment of an object if dropping it might indirectly have a
visible side effect
make `rustc_attr_parsing` less dominant in the rustc crate graph
It has/had a glob re-export of `rustc_attr_data_structures`, which is a crate much lower in the graph, and a lot of crates were using it *just* (or *mostly*) for that re-export, while they can rely on `rustc_attr_data_structures` directly.
Previous graph:
Graph with this PR:
The first commit keeps the re-export, and just changes the dependency if possible. The second commit is the "breaking change" which removes the re-export, and "explicitly" adds the `rustc_attr_data_structures` dependency where needed. It also switches over some src/tools/*.
The second commit is actually a lot more involved than I expected. Please let me know if it's a better idea to back it out and just keep the first commit.
Do not attempt to fetch a snippet from expansion. Without this change,
the inside of macros could[*] be shown as the source of the problem.
[*] Due to the way the source code is processed and reparsed in this
macro, the declarative macro has to be located outside the current
source file for the bug to appear. Otherwise, the macro call itself will
be (mis)identified as a potential `struct` field definition and the lint
will not trigger.
changelog: [`empty_struct_with_brackets`]: do not lint code coming from
macro expansion
A representation attribute `#[repr(…)]` might indicate that the ordering
of the fields or the variants is dictated by the API the code is
interfacing with. Better not lint with `arbitrary_source_item_ordering`
in this case.
changelog: [`arbitrary_source_item_ordering`]: do not lint inside items
with `#[repr]` attribute
fixes#12979fixes#12951fixes#13233
Tests still need to be finished and the docs still need to be updated,
but this should otherwise ready.
changelog: Lint `declare_interior_mutable_const` and
`borrow_interior_mutable_const` more precisely
Jason Newcomb