First commit treats all constants containing a macro call as non-local
and renames `eval_simple` to be a little clearer.
Second commit removes `CoreConstant` and treats most of them as though
they were local. A couple of the constants like `usize::MAX` are treated
as non-local as different targets may have different values.
`const_is_empty` will now ignore non-local constants since there's no
guarantee that they are the same across all targets/features/versions.
The third commit just changes some `eval` calls to `eval_local`. Most of
these were style lints which shouldn't be assuming the value of a
constant won't ever change.
changelog: none
This might be due to a low edition (< 2024) or too low a MSRV. In this
case, we will suggest only `match`.
Fixesrust-lang/rust-clippy#15744
changelog: [`unnecessary_unwrap`]: do not suggest using `if let` chains
if this is not supported with the current edition or MSRV
changelog:[`collapsible_if`]: Do not suggest using `if let` if this is
not supported with the current edition or MSRV
Make `AssocItem` aware of its impl kind
The general goal is to have fewer query dependencies by making `AssocItem` aware of its parent impl kind (inherent vs. trait) without having to query the parent def_kind.
See individual commits.
- give `ty::walk_ptrs_hir_ty` a more standard name
`peel_hir_ty_refs_and_ptrs`
- move it out of `ty`, since that's for `middle::ty::Ty`
- remove `ty::walk_ptrs_ty_depth` for equivalent `peel_middle_ty_refs`
- rename the latter into `ty::peel_and_count_ty_refs`
- incorporate mutability tracking (from
`ty::peel_mid_ty_refs_is_mutable`) into that
combine two similar arms
use in `eager_transmute`
use in `double_ended_iterator_last`
use different numbers in the new test case to avoid possible confusion
move the other "unfixable" case as well; it shouldn't lint anyway, so
having it in the main test file is fine
base check
same fields different struct
reordered fields
different paths to the same struct
same for tuple structs
style: use `zip`-the-function all over the place
makes the code a bit more concise by removing the need for explicit
`.iter()`
style: move precondition checking to the match guard
the match arms above put the "sanity" checks in the guard, and call only
`check_pat` in the body. With this commit, the (tuple) struct cases
follow that convention as well. Well, almost -- I think the ident check
belongs to the second category of checks, so I put it in the body as
well
misc: use `None` in the pattern directly
this'll probably be marginally faster thanks to the equality check being
now structural
move the tests to the right file
There are many places that join path segments with `::` to produce a
string. A lot of these use `join("::")`. Many in rustdoc use
`join_with_double_colon`, and a few use `.joined("..")`. One in Clippy
uses `itertools::join`. A couple of them look for `kw::PathRoot` in the
first segment, which can be important.
This commit introduces `rustc_ast::join_path_{syms,ident}` to do the
joining for everyone. `rustc_ast` is as good a location for these as
any, being the earliest-running of the several crates with a `Path`
type. Two functions are needed because `Ident` printing is more complex
than simple `Symbol` printing.
The commit also removes `join_with_double_colon`, and
`estimate_item_path_byte_length` with it.
There are still a handful of places that join strings with "::" that are
unchanged. They are not that important: some of them are in tests, and
some of them first split a path around "::" and then rejoin with "::".
This fixes one test case where `{{root}}` shows up in an error message.
Jason Newcomb