Auto merge of #126452 - compiler-errors:raw-lifetimes, r=<try>

Implement raw lifetimes and labels (`'r#ident`)

This PR implements a syntax for raw lifetimes and labels as requested in #126335.

This PR also moves keyword lifetime validation (i.e. checking that we don't have something like `'async`) from ast validation to the parser, similar to how we handle raw identifiers, since we don't distinguish raw identifiers after parsing.

This PR additionally extends the `keyword_idents_2024` lint to also check identifiers.

cc `@traviscross`
r? parser

Author: bors

compiler/rustc_ast/src/mut_visit.rs

@@ -776,7 +776,7 @@ fn visit_lazy_tts<T: MutVisitor>(lazy_tts: &mut Option<LazyAttrTokenStream>, vis
 pub fn visit_token<T: MutVisitor>(t: &mut Token, vis: &mut T) {
     let Token { kind, span } = t;
     match kind {
-        token::Ident(name, _) | token::Lifetime(name) => {
+        token::Ident(name, _) | token::Lifetime(name, _) => {
             let mut ident = Ident::new(*name, *span);
             vis.visit_ident(&mut ident);
             *name = ident.name;
@@ -786,7 +786,7 @@ pub fn visit_token<T: MutVisitor>(t: &mut Token, vis: &mut T) {
         token::NtIdent(ident, _is_raw) => {
             vis.visit_ident(ident);
         }
-        token::NtLifetime(ident) => {
+        token::NtLifetime(ident, _is_raw) => {
             vis.visit_ident(ident);
         }
         token::Interpolated(nt) => {

compiler/rustc_ast/src/token.rs

@@ -328,11 +328,11 @@ pub enum TokenKind {
     /// Do not forget about `NtLifetime` when you want to match on lifetime identifiers.
     /// It's recommended to use `Token::(lifetime,uninterpolate,uninterpolated_span)` to
     /// treat regular and interpolated lifetime identifiers in the same way.
-    Lifetime(Symbol),
+    Lifetime(Symbol, IdentIsRaw),
     /// This identifier (and its span) is the lifetime passed to the
     /// declarative macro. The span in the surrounding `Token` is the span of
     /// the `lifetime` metavariable in the macro's RHS.
-    NtLifetime(Ident),
+    NtLifetime(Ident, IdentIsRaw),
 
     /// An embedded AST node, as produced by a macro. This only exists for
     /// historical reasons. We'd like to get rid of it, for multiple reasons.
@@ -455,7 +455,7 @@ impl Token {
     /// if they keep spans or perform edition checks.
     pub fn uninterpolated_span(&self) -> Span {
         match self.kind {
-            NtIdent(ident, _) | NtLifetime(ident) => ident.span,
+            NtIdent(ident, _) | NtLifetime(ident, _) => ident.span,
             Interpolated(ref nt) => nt.use_span(),
             _ => self.span,
         }
@@ -627,7 +627,9 @@ impl Token {
     pub fn uninterpolate(&self) -> Cow<'_, Token> {
         match self.kind {
             NtIdent(ident, is_raw) => Cow::Owned(Token::new(Ident(ident.name, is_raw), ident.span)),
-            NtLifetime(ident) => Cow::Owned(Token::new(Lifetime(ident.name), ident.span)),
+            NtLifetime(ident, is_raw) => {
+                Cow::Owned(Token::new(Lifetime(ident.name, is_raw), ident.span))
+            }
             _ => Cow::Borrowed(self),
         }
     }
@@ -645,11 +647,11 @@ impl Token {
 
     /// Returns a lifetime identifier if this token is a lifetime.
     #[inline]
-    pub fn lifetime(&self) -> Option<Ident> {
+    pub fn lifetime(&self) -> Option<(Ident, IdentIsRaw)> {
         // We avoid using `Token::uninterpolate` here because it's slow.
         match self.kind {
-            Lifetime(name) => Some(Ident::new(name, self.span)),
-            NtLifetime(ident) => Some(ident),
+            Lifetime(name, is_raw) => Some((Ident::new(name, self.span), is_raw)),
+            NtLifetime(ident, is_raw) => Some((ident, is_raw)),
             _ => None,
         }
     }
@@ -660,6 +662,10 @@ impl Token {
     }
 
     /// Returns `true` if the token is a lifetime.
+    ///
+    /// This does not check that the lifetime is a *valid* lifetime -- for that,
+    /// you need to go through `expect_lifetime`, since we need to check that the
+    /// lifetime name is not reserved.
     pub fn is_lifetime(&self) -> bool {
         self.lifetime().is_some()
     }
@@ -832,7 +838,7 @@ impl Token {
                 _ => return None,
             },
             SingleQuote => match joint.kind {
-                Ident(name, IdentIsRaw::No) => Lifetime(Symbol::intern(&format!("'{name}"))),
+                Ident(name, is_raw) => Lifetime(Symbol::intern(&format!("'{name}")), is_raw),
                 _ => return None,
             },
 

compiler/rustc_ast/src/tokenstream.rs

@@ -487,11 +487,11 @@ impl TokenStream {
             token::NtIdent(ident, is_raw) => {
                 TokenTree::Token(Token::new(token::Ident(ident.name, is_raw), ident.span), spacing)
             }
-            token::NtLifetime(ident) => TokenTree::Delimited(
+            token::NtLifetime(ident, is_raw) => TokenTree::Delimited(
                 DelimSpan::from_single(token.span),
                 DelimSpacing::new(Spacing::JointHidden, spacing),
                 Delimiter::Invisible,
-                TokenStream::token_alone(token::Lifetime(ident.name), ident.span),
+                TokenStream::token_alone(token::Lifetime(ident.name, is_raw), ident.span),
             ),
             token::Interpolated(ref nt) => TokenTree::Delimited(
                 DelimSpan::from_single(token.span),

compiler/rustc_ast_passes/messages.ftl

@@ -156,9 +156,6 @@ ast_passes_inherent_cannot_be = inherent impls cannot be {$annotation}
     .type = inherent impl for this type
     .only_trait = only trait implementations may be annotated with {$annotation}
 
-ast_passes_invalid_label =
-    invalid label name `{$name}`
-
 ast_passes_invalid_unnamed_field =
     unnamed fields are not allowed outside of structs or unions
     .label = unnamed field declared here
@@ -170,9 +167,6 @@ ast_passes_invalid_unnamed_field_ty =
 ast_passes_item_underscore = `{$kind}` items in this context need a name
     .label = `_` is not a valid name for this `{$kind}` item
 
-ast_passes_keyword_lifetime =
-    lifetimes cannot use keyword names
-
 ast_passes_match_arm_with_no_body =
     `match` arm with no body
     .suggestion = add a body after the pattern

compiler/rustc_ast_passes/src/ast_validation.rs

@@ -257,19 +257,6 @@ impl<'a> AstValidator<'a> {
         self.session.dcx()
     }
 
-    fn check_lifetime(&self, ident: Ident) {
-        let valid_names = [kw::UnderscoreLifetime, kw::StaticLifetime, kw::Empty];
-        if !valid_names.contains(&ident.name) && ident.without_first_quote().is_reserved() {
-            self.dcx().emit_err(errors::KeywordLifetime { span: ident.span });
-        }
-    }
-
-    fn check_label(&self, ident: Ident) {
-        if ident.without_first_quote().is_reserved() {
-            self.dcx().emit_err(errors::InvalidLabel { span: ident.span, name: ident.name });
-        }
-    }
-
     fn visibility_not_permitted(&self, vis: &Visibility, note: errors::VisibilityNotPermittedNote) {
         if let VisibilityKind::Inherited = vis.kind {
             return;
@@ -879,16 +866,6 @@ impl<'a> Visitor<'a> for AstValidator<'a> {
         self.walk_ty(ty)
     }
 
-    fn visit_label(&mut self, label: &'a Label) {
-        self.check_label(label.ident);
-        visit::walk_label(self, label);
-    }
-
-    fn visit_lifetime(&mut self, lifetime: &'a Lifetime, _: visit::LifetimeCtxt) {
-        self.check_lifetime(lifetime.ident);
-        visit::walk_lifetime(self, lifetime);
-    }
-
     fn visit_field_def(&mut self, field: &'a FieldDef) {
         self.deny_unnamed_field(field);
         visit::walk_field_def(self, field)
@@ -1315,9 +1292,6 @@ impl<'a> Visitor<'a> for AstValidator<'a> {
     }
 
     fn visit_generic_param(&mut self, param: &'a GenericParam) {
-        if let GenericParamKind::Lifetime { .. } = param.kind {
-            self.check_lifetime(param.ident);
-        }
         visit::walk_generic_param(self, param);
     }
 

compiler/rustc_ast_passes/src/errors.rs

@@ -9,21 +9,6 @@ use rustc_span::{symbol::Ident, Span, Symbol};
 
 use crate::fluent_generated as fluent;
 
-#[derive(Diagnostic)]
-#[diag(ast_passes_keyword_lifetime)]
-pub struct KeywordLifetime {
-    #[primary_span]
-    pub span: Span,
-}
-
-#[derive(Diagnostic)]
-#[diag(ast_passes_invalid_label)]
-pub struct InvalidLabel {
-    #[primary_span]
-    pub span: Span,
-    pub name: Symbol,
-}
-
 #[derive(Diagnostic)]
 #[diag(ast_passes_visibility_not_permitted, code = E0449)]
 pub struct VisibilityNotPermitted {

compiler/rustc_ast_pretty/src/pprust/state.rs

@@ -12,7 +12,9 @@ use crate::pprust::state::fixup::FixupContext;
 use ast::TraitBoundModifiers;
 use rustc_ast::attr::AttrIdGenerator;
 use rustc_ast::ptr::P;
-use rustc_ast::token::{self, BinOpToken, CommentKind, Delimiter, Nonterminal, Token, TokenKind};
+use rustc_ast::token::{
+    self, BinOpToken, CommentKind, Delimiter, IdentIsRaw, Nonterminal, Token, TokenKind,
+};
 use rustc_ast::tokenstream::{Spacing, TokenStream, TokenTree};
 use rustc_ast::util::classify;
 use rustc_ast::util::comments::{Comment, CommentStyle};
@@ -946,8 +948,13 @@ pub trait PrintState<'a>: std::ops::Deref<Target = pp::Printer> + std::ops::Dere
             token::NtIdent(ident, is_raw) => {
                 IdentPrinter::for_ast_ident(ident, is_raw.into()).to_string().into()
             }
-            token::Lifetime(name) => name.to_string().into(),
-            token::NtLifetime(ident) => ident.name.to_string().into(),
+
+            token::Lifetime(name, IdentIsRaw::No)
+            | token::NtLifetime(Ident { name, .. }, IdentIsRaw::No) => name.to_string().into(),
+            token::Lifetime(name, IdentIsRaw::Yes)
+            | token::NtLifetime(Ident { name, .. }, IdentIsRaw::Yes) => {
+                format!("'r#{}", &name.as_str()[1..]).into()
+            }
 
             /* Other */
             token::DocComment(comment_kind, attr_style, data) => {

compiler/rustc_expand/src/mbe/macro_parser.rs

@@ -398,8 +398,10 @@ pub(crate) enum NamedMatch {
 fn token_name_eq(t1: &Token, t2: &Token) -> bool {
     if let (Some((ident1, is_raw1)), Some((ident2, is_raw2))) = (t1.ident(), t2.ident()) {
         ident1.name == ident2.name && is_raw1 == is_raw2
-    } else if let (Some(ident1), Some(ident2)) = (t1.lifetime(), t2.lifetime()) {
-        ident1.name == ident2.name
+    } else if let (Some((ident1, is_raw1)), Some((ident2, is_raw2))) =
+        (t1.lifetime(), t2.lifetime())
+    {
+        ident1.name == ident2.name && is_raw1 == is_raw2
     } else {
         t1.kind == t2.kind
     }

compiler/rustc_expand/src/mbe/transcribe.rs

@@ -278,9 +278,9 @@ pub(super) fn transcribe<'a>(
                             let kind = token::NtIdent(*ident, *is_raw);
                             TokenTree::token_alone(kind, sp)
                         }
-                        MatchedSingle(ParseNtResult::Lifetime(ident)) => {
+                        MatchedSingle(ParseNtResult::Lifetime(ident, is_raw)) => {
                             marker.visit_span(&mut sp);
-                            let kind = token::NtLifetime(*ident);
+                            let kind = token::NtLifetime(*ident, *is_raw);
                             TokenTree::token_alone(kind, sp)
                         }
                         MatchedSingle(ParseNtResult::Nt(nt)) => {

compiler/rustc_expand/src/proc_macro_server.rs

@@ -227,15 +227,16 @@ impl FromInternal<(TokenStream, &mut Rustc<'_, '_>)> for Vec<TokenTree<TokenStre
                     span: ident.span,
                 })),
 
-                Lifetime(name) => {
+                Lifetime(name, is_raw) => {
                     let ident = symbol::Ident::new(name, span).without_first_quote();
                     trees.extend([
                         TokenTree::Punct(Punct { ch: b'\'', joint: true, span }),
-                        TokenTree::Ident(Ident { sym: ident.name, is_raw: false, span }),
+                        TokenTree::Ident(Ident { sym: ident.name, is_raw: is_raw.into(), span }),
                     ]);
                 }
-                NtLifetime(ident) => {
-                    let stream = TokenStream::token_alone(token::Lifetime(ident.name), ident.span);
+                NtLifetime(ident, is_raw) => {
+                    let stream =
+                        TokenStream::token_alone(token::Lifetime(ident.name, is_raw), ident.span);
                     trees.push(TokenTree::Group(Group {
                         delimiter: pm::Delimiter::None,
                         stream: Some(stream),

compiler/rustc_lexer/src/lib.rs

@@ -104,6 +104,9 @@ pub enum TokenKind {
     /// "'a"
     Lifetime { starts_with_number: bool },
 
+    /// `'r#async`
+    RawLifetime,
+
     // One-char tokens:
     /// ";"
     Semi,
@@ -676,9 +679,18 @@ impl Cursor<'_> {
             return Literal { kind, suffix_start };
         }
 
+        if self.first() == 'r' && self.second() == '#' && is_id_start(self.third()) {
+            // Eat "r" character.
+            self.bump();
+            // Eat "#" symbol.
+            self.bump();
+            // Eat the identifier part of RawIdent.
+            self.eat_identifier();
+            return RawLifetime;
+        }
+
         // Either a lifetime or a character literal with
         // length greater than 1.
-
         let starts_with_number = self.first().is_ascii_digit();
 
         // Skip the literal contents.

compiler/rustc_lint/src/builtin.rs

@@ -1888,8 +1888,16 @@ impl KeywordIdents {
                 TokenTree::Token(token, _) => {
                     if let Some((ident, token::IdentIsRaw::No)) = token.ident() {
                         if !prev_dollar {
-                            self.check_ident_token(cx, UnderMacro(true), ident);
+                            self.check_ident_token(cx, UnderMacro(true), ident, "");
                         }
+                    }
+                    if let Some((ident, token::IdentIsRaw::No)) = token.lifetime() {
+                        self.check_ident_token(
+                            cx,
+                            UnderMacro(true),
+                            ident.without_first_quote(),
+                            "'",
+                        );
                     } else if token.kind == TokenKind::Dollar {
                         prev_dollar = true;
                         continue;
@@ -1906,6 +1914,7 @@ impl KeywordIdents {
         cx: &EarlyContext<'_>,
         UnderMacro(under_macro): UnderMacro,
         ident: Ident,
+        prefix: &'static str,
     ) {
         let (lint, edition) = match ident.name {
             kw::Async | kw::Await | kw::Try => (KEYWORD_IDENTS_2018, Edition::Edition2018),
@@ -1939,7 +1948,7 @@ impl KeywordIdents {
         cx.emit_span_lint(
             lint,
             ident.span,
-            BuiltinKeywordIdents { kw: ident, next: edition, suggestion: ident.span },
+            BuiltinKeywordIdents { kw: ident, next: edition, suggestion: ident.span, prefix },
         );
     }
 }
@@ -1952,7 +1961,11 @@ impl EarlyLintPass for KeywordIdents {
         self.check_tokens(cx, &mac.args.tokens);
     }
     fn check_ident(&mut self, cx: &EarlyContext<'_>, ident: Ident) {
-        self.check_ident_token(cx, UnderMacro(false), ident);
+        if ident.name.as_str().starts_with('\'') {
+            self.check_ident_token(cx, UnderMacro(false), ident.without_first_quote(), "'");
+        } else {
+            self.check_ident_token(cx, UnderMacro(false), ident, "");
+        }
     }
 }
 

compiler/rustc_lint/src/lints.rs

@@ -368,8 +368,9 @@ pub enum BuiltinEllipsisInclusiveRangePatternsLint {
 pub struct BuiltinKeywordIdents {
     pub kw: Ident,
     pub next: Edition,
-    #[suggestion(code = "r#{kw}", applicability = "machine-applicable")]
+    #[suggestion(code = "{prefix}r#{kw}", applicability = "machine-applicable")]
     pub suggestion: Span,
+    pub prefix: &'static str,
 }
 
 #[derive(LintDiagnostic)]

compiler/rustc_parse/messages.ftl

@@ -390,6 +390,9 @@ parse_invalid_dyn_keyword = invalid `dyn` keyword
 parse_invalid_expression_in_let_else = a `{$operator}` expression cannot be directly assigned in `let...else`
 parse_invalid_identifier_with_leading_number = identifiers cannot start with a number
 
+parse_invalid_label =
+    invalid label name `{$name}`
+
 parse_invalid_literal_suffix_on_tuple_index = suffixes on a tuple index are invalid
     .label = invalid suffix `{$suffix}`
     .tuple_exception_line_1 = `{$suffix}` is *temporarily* accepted on tuple index fields as it was incorrectly accepted on stable for a few releases
@@ -416,6 +419,9 @@ parse_invalid_unicode_escape = invalid unicode character escape
 parse_invalid_variable_declaration =
     invalid variable declaration
 
+parse_keyword_lifetime =
+    lifetimes cannot use keyword names
+
 parse_kw_bad_case = keyword `{$kw}` is written in the wrong case
     .suggestion = write it in the correct case
 

compiler/rustc_parse/src/errors.rs

@@ -1907,6 +1907,21 @@ pub struct CannotBeRawIdent {
     pub ident: Symbol,
 }
 
+#[derive(Diagnostic)]
+#[diag(parse_keyword_lifetime)]
+pub struct KeywordLifetime {
+    #[primary_span]
+    pub span: Span,
+}
+
+#[derive(Diagnostic)]
+#[diag(parse_invalid_label)]
+pub struct InvalidLabel {
+    #[primary_span]
+    pub span: Span,
+    pub name: Symbol,
+}
+
 #[derive(Diagnostic)]
 #[diag(parse_cr_doc_comment)]
 pub struct CrDocComment {