Extend doc keyword feature by allowing any ident
Part of #51315.
As suggested by ``@danielhenrymantilla`` in [this comment](https://github.com/rust-lang/rust/issues/51315#issuecomment-733879934), this PR extends `#[doc(keyword = "...")]` to allow any ident to be used as keyword. The final goal is to allow (proc-)macro crates' owners to write documentation of the keywords they might introduce.
r? ``@jyn514``
Rename "stability" CSS class to "item-info" and combine `document_stability` with `document_short`
Follow-up of #79300
The point of this PR is to make the CSS class more accurate since it's not only about stability anymore.
r? ``@jyn514``
Properly handle attributes on statements
We now collect tokens for the underlying node wrapped by `StmtKind`
nstead of storing tokens directly in `Stmt`.
`LazyTokenStream` now supports capturing a trailing semicolon after it
is initially constructed. This allows us to avoid refactoring statement
parsing to wrap the parsing of the semicolon in `parse_tokens`.
Attributes on item statements
(e.g. `fn foo() { #[bar] struct MyStruct; }`) are now treated as
item attributes, not statement attributes, which is consistent with how
we handle attributes on other kinds of statements. The feature-gating
code is adjusted so that proc-macro attributes are still allowed on item
statements on stable.
Two built-in macros (`#[global_allocator]` and `#[test]`) needed to be
adjusted to support being passed `Annotatable::Stmt`.
For example, `1` is parsed as an integer literal, but it can be turned
into a float with the suffix `f32`. Now the error calls them "numeric
literals" and notes that you can add a float suffix since they can be
either integers or floats.
Cleanup more of rustdoc
- Use `Item::from_def_id` for StructField
- Use `from_def_id_and_parts` for primitives and keywords
- Take `String` instead of `Symbol` in `from_def_id` - this avoids having to intern then immediately stringify the existing string.
- Remove unused `get_stability` and `get_deprecation`
- Remove unused `attrs` field from `primitives`
- Remove unused `attrs` field from `keywords`
This will probably conflict with https://github.com/rust-lang/rust/pull/79335 and I would prefer for that PR to land first - I'm anxious for https://github.com/rust-lang/rust/pull/77467 to land :)
Makes https://github.com/rust-lang/rust/issues/76998 easier to add.
r? `@GuillaumeGomez`
This preserves the current lint behavior for now.
Linting after item statements currently prevents the compiler from bootstrapping.
Fixing this is blocked on fixing this upstream in Cargo, and bumping the Cargo
submodule.
Loading a macro from libstd causes us to load serialized
`SyntaxContext`s in a platform-dependent way, causing the printed spans
to differ between platforms.
When parsing a statement (e.g. inside a function body),
we now consider `struct Foo {};` and `$stmt;` to each consist
of two statements: `struct Foo {}` and `;`, and `$stmt` and `;`.
As a result, an attribute macro invoke as
`fn foo() { #[attr] struct Bar{}; }` will see `struct Bar{}` as its
input. Additionally, the 'unused semicolon' lint now fires in more
places.
We now collect tokens for the underlying node wrapped by `StmtKind`
instead of storing tokens directly in `Stmt`.
`LazyTokenStream` now supports capturing a trailing semicolon after it
is initially constructed. This allows us to avoid refactoring statement
parsing to wrap the parsing of the semicolon in `parse_tokens`.
Attributes on item statements
(e.g. `fn foo() { #[bar] struct MyStruct; }`) are now treated as
item attributes, not statement attributes, which is consistent with how
we handle attributes on other kinds of statements. The feature-gating
code is adjusted so that proc-macro attributes are still allowed on item
statements on stable.
Two built-in macros (`#[global_allocator]` and `#[test]`) needed to be
adjusted to support being passed `Annotatable::Stmt`.
Cache pretty-print/retokenize result to avoid compile time blowup
Fixes#79242
If a `macro_rules!` recursively builds up a nested nonterminal
(passing it to a proc-macro at each step), we will end up repeatedly
pretty-printing/retokenizing the same nonterminals. Unfortunately, the
'probable equality' check we do has a non-trivial cost, which leads to a
blowup in compilation time.
As a workaround, we cache the result of the 'probable equality' check,
which eliminates the compilation time blowup for the linked issue. This
commit only touches a single file (other than adding tests), so it
should be easy to backport.
The proper solution is to remove the pretty-print/retokenize hack
entirely. However, this will almost certainly break a large number of
crates that were relying on hygiene bugs created by using the reparsed
`TokenStream`. As a result, we will definitely not want to backport
such a change.
Normalize `<X as Y>::T` for rustdoc
- Only run for `QPath::Resolved` with `Some` self parameter (`<X as Y>::T`)
- Fall back to the previous behavior if the path can't be resolved
The first commit is a pure refactor and should probably be reviewed by `@GuillaumeGomez.` I recommend reviewing the second commit on its own.
Fixes https://github.com/rust-lang/rust/issues/77459.
r? `@eddyb`
cc `@danielhenrymantilla` , `@lcnr`
