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.
name resolution for guard patterns
This PR provides an initial implementation of name resolution for guard patterns [(RFC 3637)](https://github.com/rust-lang/rfcs/blob/master/text/3637-guard-patterns.md). This does not change the requirement that the bindings on either side of an or-pattern must be the same [(proposal here)](https://github.com/rust-lang/rfcs/blob/master/text/3637-guard-patterns.md#allowing-mismatching-bindings-when-possible); the code that handles that is separate from what this PR touches, so I'm saving it for a follow-up.
On a technical level, this separates "collecting the bindings in a pattern" (which was already done for or-patterns) from "introducing those bindings into scope". I believe the approach used here can be extended straightforwardly in the future to work with `if let` guard patterns, but I haven't tried it myself since we don't allow those yet.
Tracking issue for guard patterns: #129967
cc ``@Nadrieril``
Prefer to suggest stable candidates rather than unstable ones
Fixes#140240
The logic is to replace unstable suggestions if we meet a new stable one, and do nothing if any other situation. In old logic, we just use the first candidate we meet as the suggestion for the same items.
E.g., `std::range::legacy::Range` vs `std::ops::Range`, `legacy` in the former is unstable, we prefer to suggest use the latter.
All uses have been removed. And it's nonsensical: an identifier by
definition has at least one char.
The commits adds an is-non-empty assertion to `Ident::new` to enforce
this, and converts some `Ident` constructions to use `Ident::new`.
Adding the assertion requires making `Ident::new` and
`Ident::with_dummy_span` non-const, which is no great loss.
The commit amends a couple of places that do path splitting to ensure no
empty identifiers are created.
This splits introduction of bindings into scope
(`apply_pattern_bindings`) apart from manipulation of the pattern's
binding map (`fresh_binding`). By delaying the latter, we can keep
bindings from appearing in-scope in guards.
Since `fresh_binding` is now specifically for manipulating a pattern's
bindings map, this commit also inlines a use of `fresh_binding` that was
only adding to the innermost rib.
I'll be modifying it in future commits, so I think it's cleanest to
abstract it out. Possibly a newtype would be ideal, but for now this is
least disruptive.
Remove global `next_disambiguator` state and handle it with a `DisambiguatorState` type
This removes `Definitions.next_disambiguator` as it doesn't guarantee deterministic def paths when `create_def` is called in parallel. Instead a new `DisambiguatorState` type is passed as a mutable reference to `create_def` to help create unique def paths. `create_def` calls with distinct `DisambiguatorState` instances must ensure that that the def paths are unique without its help.
Anon associated types did rely on this global state for uniqueness and are changed to use (method they're defined in + their position in the method return type) as the `DefPathData` to ensure uniqueness. This also means that the method they're defined in appears in error messages, which is nicer.
`DefPathData::NestedStatic` is added to use for nested data inside statics instead of reusing `DefPathData::AnonConst` to avoid conflicts with those.
cc `@oli-obk`
Simplify `LazyAttrTokenStream`
`LazyAttrTokenStream` is an unpleasant type: `Lrc<Box<dyn ToAttrTokenStream>>`. Why does it look like that?
- There are two `ToAttrTokenStream` impls, one for the lazy case, and one for the case where we already have an `AttrTokenStream`.
- The lazy case (`LazyAttrTokenStreamImpl`) is implemented in `rustc_parse`, but `LazyAttrTokenStream` is defined in `rustc_ast`, which does not depend on `rustc_parse`. The use of the trait lets `rustc_ast` implicitly depend on `rustc_parse`. This explains the `dyn`.
- `LazyAttrTokenStream` must have a `size_of` as small as possible, because it's used in many AST nodes. This explains the `Lrc<Box<_>>`, which keeps it to one word. (It's required `Lrc<dyn _>` would be a fat pointer.)
This PR moves `LazyAttrTokenStreamImpl` (and a few other token stream things) from `rustc_parse` to `rustc_ast`. This lets us replace the `ToAttrTokenStream` trait with a two-variant enum and also remove the `Box`, changing `LazyAttrTokenStream` to `Lrc<LazyAttrTokenStreamInner>`. Plus it does a few cleanups.
r? `@petrochenkov`
This commit does the following.
- Changes it from `Lrc<Box<dyn ToAttrTokenStream>>` to
`Lrc<LazyAttrTokenStreamInner>`.
- Reworks `LazyAttrTokenStreamImpl` as `LazyAttrTokenStreamInner`, which
is a two-variant enum.
- Removes the `ToAttrTokenStream` trait and the two impls of it.
The recursion limit must be increased in some crates otherwise rustdoc
aborts.
hygiene: Rename semi-transparent to semi-opaque
"Semi-transparent" is just too damn long for a name, especially when used multiple times on a single line, it bothered me when working on #139083.
An optimist sees a macro as semi-opaque, a pessimist sees it as semi-transparent.
Or is it the other way round?
This is pretty weird code. As the `HACK` comment indicates, we push the
empty ident here only to make the path longer, so certain checks to
occur within `lint_if_path_starts_with_module`. `dummy` is a better
choice because it explicitly communicates that the actual value doesn't
matter.
Overhaul `AssocItem`
`AssocItem` has multiple fields that only make sense some of the time. E.g. the `name` can be empty if it's an RPITIT associated type. It's clearer and less error prone if these fields are moved to the relevant `kind` variants.
r? ``@fee1-dead``
To accurately reflect that RPITIT assoc items don't have a name. This
avoids the use of `kw::Empty` to mean "no name", which is error prone.
Helps with #137978.
Fix up partial res of segment in primitive resolution hack
There is a hack in the resolver:
```
// In `a(::assoc_item)*` `a` cannot be a module. If `a` does resolve to a module we
// don't report an error right away, but try to fallback to a primitive type.
```
This fixes up the resolution for primitives which would otherwise resolve to a module, but we weren't also updating the res of the path segment, leading to weird diagnostics.
We explicitly call `self.r.partial_res_map.insert` instead of `record_partial_res` b/c we have recorded a partial res already, and we specifically want to override it.
cc https://github.com/rust-lang/rust/issues/139095#issuecomment-2764371934
`hir::AssocItem` currently has a boolean `fn_has_self_parameter` field,
which is misplaced, because it's only relevant for associated fns, not
for associated consts or types. This commit moves it (and renames it) to
the `AssocKind::Fn` variant, where it belongs.
This requires introducing a new C-style enum, `AssocTag`, which is like
`AssocKind` but without the fields. This is because `AssocKind` values
are passed to various functions like `find_by_ident_and_kind` to
indicate what kind of associated item should be searched for, and having
to specify `has_self` isn't relevant there.
New methods:
- Predicates `AssocItem::is_fn` and `AssocItem::is_method`.
- `AssocItem::as_tag` which converts `AssocItem::kind` to `AssocTag`.
Removed `find_by_name_and_kinds`, which is unused.
`AssocItem::descr` can now distinguish between methods and associated
functions, which slightly improves some error messages.
Rename some `name` variables as `ident`.
It bugs me when variables of type `Ident` are called `name`. It leads to silly things like `name.name`. `Ident` variables should be called `ident`, and `name` should be used for variables of type `Symbol`.
This commit improves things by by doing `s/name/ident/` on a bunch of `Ident` variables. Not all of them, but a decent chunk.
r? `@fee1-dead`
It bugs me when variables of type `Ident` are called `name`. It leads to
silly things like `name.name`. `Ident` variables should be called
`ident`, and `name` should be used for variables of type `Symbol`.
This commit improves things by by doing `s/name/ident/` on a bunch of
`Ident` variables. Not all of them, but a decent chunk.
`resolve_ident_in_lexical_scope` checks for an empty name. Why is this
necessary? Because `parse_item_impl` can produce an `impl` block with an
empty trait name in some cases. This is pretty gross and very
non-obvious.
This commit avoids the use of the empty trait name. In one case the
trait name is instead pulled from `TyKind::ImplTrait`, which prevents
the output for `tests/ui/impl-trait/extra-impl-in-trait-impl.rs` from
changing. In the other case we just fail the parse and don't try to
recover. I think losing error recovery in this obscure case is worth
the code cleanup.
This change affects `tests/ui/parser/impl-parsing.rs`, which is split in
two, and the obsolete `..` syntax cases are removed (they are tested
elsewhere).
Add new `PatKind::Missing` variants
To avoid some ugly uses of `kw::Empty` when handling "missing" patterns, e.g. in bare fn tys. Helps with #137978. Details in the individual commits.
r? ``@oli-obk``
Note potential but private items in show_candidates
Closes#138626 .
We should add potential private items to give ample hints.
And for the other seemingly false positive ` pub use crate:1️⃣:Foo;` should be kept because we don't know if the user wants to import other module's items or not, and therefore should be given the full option to do so.
r? compiler
`ast::Item` has an `ident` field.
- It's always non-empty for these item kinds: `ExternCrate`, `Static`,
`Const`, `Fn`, `Mod`, `TyAlias`, `Enum`, `Struct`, `Union`,
`Trait`, `TraitAlias`, `MacroDef`, `Delegation`.
- It's always empty for these item kinds: `Use`, `ForeignMod`,
`GlobalAsm`, `Impl`, `MacCall`, `DelegationMac`.
There is a similar story for `AssocItemKind` and `ForeignItemKind`.
Some sites that handle items check for an empty ident, some don't. This
is a very C-like way of doing things, but this is Rust, we have sum
types, we can do this properly and never forget to check for the
exceptional case and never YOLO possibly empty identifiers (or possibly
dummy spans) around and hope that things will work out.
The commit is large but it's mostly obvious plumbing work. Some notable
things.
- `ast::Item` got 8 bytes bigger. This could be avoided by boxing the
fields within some of the `ast::ItemKind` variants (specifically:
`Struct`, `Union`, `Enum`). I might do that in a follow-up; this
commit is big enough already.
- For the visitors: `FnKind` no longer needs an `ident` field because
the `Fn` within how has one.
- In the parser, the `ItemInfo` typedef is no longer needed. It was used
in various places to return an `Ident` alongside an `ItemKind`, but
now the `Ident` (if present) is within the `ItemKind`.
- In a few places I renamed identifier variables called `name` (or
`foo_name`) as `ident` (or `foo_ident`), to better match the type, and
because `name` is normally used for `Symbol`s. It's confusing to see
something like `foo_name.name`.
