Merge pull request #20128 from lnicola/sync-from-rust

Sync from downstream

10 files changed, 104 insertions, 90 deletions

diff --git a/compiler/rustc_parse/src/lexer/mod.rs b/compiler/rustc_parse/src/lexer/mod.rs
index 2845bbed1c0..60d275bf2b4 100644
--- a/compiler/rustc_parse/src/lexer/mod.rs
+++ b/compiler/rustc_parse/src/lexer/mod.rs
@@ -1,5 +1,3 @@
-use std::ops::Range;
-
 use diagnostics::make_unclosed_delims_error;
 use rustc_ast::ast::{self, AttrStyle};
 use rustc_ast::token::{self, CommentKind, Delimiter, IdentIsRaw, Token, TokenKind};
@@ -10,7 +8,7 @@ use rustc_errors::{Applicability, Diag, DiagCtxtHandle, StashKey};
 use rustc_lexer::{
     Base, Cursor, DocStyle, FrontmatterAllowed, LiteralKind, RawStrError, is_whitespace,
 };
-use rustc_literal_escaper::{EscapeError, Mode, unescape_mixed, unescape_unicode};
+use rustc_literal_escaper::{EscapeError, Mode, check_for_errors};
 use rustc_session::lint::BuiltinLintDiag;
 use rustc_session::lint::builtin::{
     RUST_2021_PREFIXES_INCOMPATIBLE_SYNTAX, RUST_2024_GUARDED_STRING_INCOMPATIBLE_SYNTAX,
@@ -702,7 +700,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
                     }
                     err.emit()
                 }
-                self.cook_unicode(token::Char, Mode::Char, start, end, 1, 1) // ' '
+                self.cook_quoted(token::Char, Mode::Char, start, end, 1, 1) // ' '
             }
             rustc_lexer::LiteralKind::Byte { terminated } => {
                 if !terminated {
@@ -714,7 +712,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
                         .with_code(E0763)
                         .emit()
                 }
-                self.cook_unicode(token::Byte, Mode::Byte, start, end, 2, 1) // b' '
+                self.cook_quoted(token::Byte, Mode::Byte, start, end, 2, 1) // b' '
             }
             rustc_lexer::LiteralKind::Str { terminated } => {
                 if !terminated {
@@ -726,7 +724,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
                         .with_code(E0765)
                         .emit()
                 }
-                self.cook_unicode(token::Str, Mode::Str, start, end, 1, 1) // " "
+                self.cook_quoted(token::Str, Mode::Str, start, end, 1, 1) // " "
             }
             rustc_lexer::LiteralKind::ByteStr { terminated } => {
                 if !terminated {
@@ -738,7 +736,8 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
                         .with_code(E0766)
                         .emit()
                 }
-                self.cook_unicode(token::ByteStr, Mode::ByteStr, start, end, 2, 1) // b" "
+                self.cook_quoted(token::ByteStr, Mode::ByteStr, start, end, 2, 1)
+                // b" "
             }
             rustc_lexer::LiteralKind::CStr { terminated } => {
                 if !terminated {
@@ -750,13 +749,14 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
                         .with_code(E0767)
                         .emit()
                 }
-                self.cook_mixed(token::CStr, Mode::CStr, start, end, 2, 1) // c" "
+                self.cook_quoted(token::CStr, Mode::CStr, start, end, 2, 1) // c" "
             }
             rustc_lexer::LiteralKind::RawStr { n_hashes } => {
                 if let Some(n_hashes) = n_hashes {
                     let n = u32::from(n_hashes);
                     let kind = token::StrRaw(n_hashes);
-                    self.cook_unicode(kind, Mode::RawStr, start, end, 2 + n, 1 + n) // r##" "##
+                    self.cook_quoted(kind, Mode::RawStr, start, end, 2 + n, 1 + n)
+                // r##" "##
                 } else {
                     self.report_raw_str_error(start, 1);
                 }
@@ -765,7 +765,8 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
                 if let Some(n_hashes) = n_hashes {
                     let n = u32::from(n_hashes);
                     let kind = token::ByteStrRaw(n_hashes);
-                    self.cook_unicode(kind, Mode::RawByteStr, start, end, 3 + n, 1 + n) // br##" "##
+                    self.cook_quoted(kind, Mode::RawByteStr, start, end, 3 + n, 1 + n)
+                // br##" "##
                 } else {
                     self.report_raw_str_error(start, 2);
                 }
@@ -774,7 +775,8 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
                 if let Some(n_hashes) = n_hashes {
                     let n = u32::from(n_hashes);
                     let kind = token::CStrRaw(n_hashes);
-                    self.cook_unicode(kind, Mode::RawCStr, start, end, 3 + n, 1 + n) // cr##" "##
+                    self.cook_quoted(kind, Mode::RawCStr, start, end, 3 + n, 1 + n)
+                // cr##" "##
                 } else {
                     self.report_raw_str_error(start, 2);
                 }
@@ -1091,7 +1093,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
         self.dcx().emit_fatal(errors::TooManyHashes { span: self.mk_sp(start, self.pos), num });
     }
 
-    fn cook_common(
+    fn cook_quoted(
         &self,
         mut kind: token::LitKind,
         mode: Mode,
@@ -1099,32 +1101,28 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
         end: BytePos,
         prefix_len: u32,
         postfix_len: u32,
-        unescape: fn(&str, Mode, &mut dyn FnMut(Range<usize>, Result<(), EscapeError>)),
     ) -> (token::LitKind, Symbol) {
         let content_start = start + BytePos(prefix_len);
         let content_end = end - BytePos(postfix_len);
         let lit_content = self.str_from_to(content_start, content_end);
-        unescape(lit_content, mode, &mut |range, result| {
-            // Here we only check for errors. The actual unescaping is done later.
-            if let Err(err) = result {
-                let span_with_quotes = self.mk_sp(start, end);
-                let (start, end) = (range.start as u32, range.end as u32);
-                let lo = content_start + BytePos(start);
-                let hi = lo + BytePos(end - start);
-                let span = self.mk_sp(lo, hi);
-                let is_fatal = err.is_fatal();
-                if let Some(guar) = emit_unescape_error(
-                    self.dcx(),
-                    lit_content,
-                    span_with_quotes,
-                    span,
-                    mode,
-                    range,
-                    err,
-                ) {
-                    assert!(is_fatal);
-                    kind = token::Err(guar);
-                }
+        check_for_errors(lit_content, mode, |range, err| {
+            let span_with_quotes = self.mk_sp(start, end);
+            let (start, end) = (range.start as u32, range.end as u32);
+            let lo = content_start + BytePos(start);
+            let hi = lo + BytePos(end - start);
+            let span = self.mk_sp(lo, hi);
+            let is_fatal = err.is_fatal();
+            if let Some(guar) = emit_unescape_error(
+                self.dcx(),
+                lit_content,
+                span_with_quotes,
+                span,
+                mode,
+                range,
+                err,
+            ) {
+                assert!(is_fatal);
+                kind = token::Err(guar);
             }
         });
 
@@ -1137,34 +1135,6 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
         };
         (kind, sym)
     }
-
-    fn cook_unicode(
-        &self,
-        kind: token::LitKind,
-        mode: Mode,
-        start: BytePos,
-        end: BytePos,
-        prefix_len: u32,
-        postfix_len: u32,
-    ) -> (token::LitKind, Symbol) {
-        self.cook_common(kind, mode, start, end, prefix_len, postfix_len, |src, mode, callback| {
-            unescape_unicode(src, mode, &mut |span, result| callback(span, result.map(drop)))
-        })
-    }
-
-    fn cook_mixed(
-        &self,
-        kind: token::LitKind,
-        mode: Mode,
-        start: BytePos,
-        end: BytePos,
-        prefix_len: u32,
-        postfix_len: u32,
-    ) -> (token::LitKind, Symbol) {
-        self.cook_common(kind, mode, start, end, prefix_len, postfix_len, |src, mode, callback| {
-            unescape_mixed(src, mode, &mut |span, result| callback(span, result.map(drop)))
-        })
-    }
 }
 
 pub fn nfc_normalize(string: &str) -> Symbol {
diff --git a/compiler/rustc_parse/src/parser/diagnostics.rs b/compiler/rustc_parse/src/parser/diagnostics.rs
index a64bb424117..1df0ccbd8af 100644
--- a/compiler/rustc_parse/src/parser/diagnostics.rs
+++ b/compiler/rustc_parse/src/parser/diagnostics.rs
@@ -2731,7 +2731,7 @@ impl<'a> Parser<'a> {
             return first_pat;
         }
         if !matches!(first_pat.kind, PatKind::Ident(_, _, None) | PatKind::Path(..))
-            || !self.look_ahead(1, |token| token.is_ident() && !token.is_reserved_ident())
+            || !self.look_ahead(1, |token| token.is_non_reserved_ident())
         {
             let mut snapshot_type = self.create_snapshot_for_diagnostic();
             snapshot_type.bump(); // `:`
diff --git a/compiler/rustc_parse/src/parser/expr.rs b/compiler/rustc_parse/src/parser/expr.rs
index 93c76c47f06..3cedc86dc0d 100644
--- a/compiler/rustc_parse/src/parser/expr.rs
+++ b/compiler/rustc_parse/src/parser/expr.rs
@@ -847,7 +847,7 @@ impl<'a> Parser<'a> {
         self.dcx().emit_err(errors::LifetimeInBorrowExpression { span, lifetime_span: lt_span });
     }
 
-    /// Parse `mut?` or `raw [ const | mut ]`.
+    /// Parse `mut?` or `[ raw | pin ] [ const | mut ]`.
     fn parse_borrow_modifiers(&mut self) -> (ast::BorrowKind, ast::Mutability) {
         if self.check_keyword(exp!(Raw)) && self.look_ahead(1, Token::is_mutability) {
             // `raw [ const | mut ]`.
@@ -855,6 +855,11 @@ impl<'a> Parser<'a> {
             assert!(found_raw);
             let mutability = self.parse_const_or_mut().unwrap();
             (ast::BorrowKind::Raw, mutability)
+        } else if let Some((ast::Pinnedness::Pinned, mutbl)) = self.parse_pin_and_mut() {
+            // `pin [ const | mut ]`.
+            // `pin` has been gated in `self.parse_pin_and_mut()` so we don't
+            // need to gate it here.
+            (ast::BorrowKind::Pin, mutbl)
         } else {
             // `mut?`
             (ast::BorrowKind::Ref, self.parse_mutability())
@@ -3870,8 +3875,7 @@ impl<'a> Parser<'a> {
             // Check if a colon exists one ahead. This means we're parsing a fieldname.
             let is_shorthand = !this.look_ahead(1, |t| t == &token::Colon || t == &token::Eq);
             // Proactively check whether parsing the field will be incorrect.
-            let is_wrong = this.token.is_ident()
-                && !this.token.is_reserved_ident()
+            let is_wrong = this.token.is_non_reserved_ident()
                 && !this.look_ahead(1, |t| {
                     t == &token::Colon
                         || t == &token::Eq
diff --git a/compiler/rustc_parse/src/parser/generics.rs b/compiler/rustc_parse/src/parser/generics.rs
index c05479feb61..af1d1a1ec66 100644
--- a/compiler/rustc_parse/src/parser/generics.rs
+++ b/compiler/rustc_parse/src/parser/generics.rs
@@ -353,6 +353,20 @@ impl<'a> Parser<'a> {
         if !self.eat_keyword(exp!(Where)) {
             return Ok((where_clause, None));
         }
+
+        if self.eat_noexpect(&token::Colon) {
+            let colon_span = self.prev_token.span;
+            self.dcx()
+                .struct_span_err(colon_span, "unexpected colon after `where`")
+                .with_span_suggestion_short(
+                    colon_span,
+                    "remove the colon",
+                    "",
+                    Applicability::MachineApplicable,
+                )
+                .emit();
+        }
+
         where_clause.has_where_token = true;
         let where_lo = self.prev_token.span;
 
diff --git a/compiler/rustc_parse/src/parser/item.rs b/compiler/rustc_parse/src/parser/item.rs
index 5088caa80f8..9ed7124a11c 100644
--- a/compiler/rustc_parse/src/parser/item.rs
+++ b/compiler/rustc_parse/src/parser/item.rs
@@ -1552,7 +1552,8 @@ impl<'a> Parser<'a> {
             })
             .map_err(|mut err| {
                 err.span_label(ident.span, "while parsing this enum");
-                if self.token == token::Colon {
+                // Try to recover `enum Foo { ident : Ty }`.
+                if self.prev_token.is_non_reserved_ident() && self.token == token::Colon {
                     let snapshot = self.create_snapshot_for_diagnostic();
                     self.bump();
                     match self.parse_ty() {
diff --git a/compiler/rustc_parse/src/parser/mod.rs b/compiler/rustc_parse/src/parser/mod.rs
index b2e90251367..cfc0399b0ca 100644
--- a/compiler/rustc_parse/src/parser/mod.rs
+++ b/compiler/rustc_parse/src/parser/mod.rs
@@ -685,7 +685,7 @@ impl<'a> Parser<'a> {
 
     /// Is the given keyword `kw` followed by a non-reserved identifier?
     fn is_kw_followed_by_ident(&self, kw: Symbol) -> bool {
-        self.token.is_keyword(kw) && self.look_ahead(1, |t| t.is_ident() && !t.is_reserved_ident())
+        self.token.is_keyword(kw) && self.look_ahead(1, |t| t.is_non_reserved_ident())
     }
 
     #[inline]
diff --git a/compiler/rustc_parse/src/parser/path.rs b/compiler/rustc_parse/src/parser/path.rs
index 9bce2fa74ca..1f4049f197f 100644
--- a/compiler/rustc_parse/src/parser/path.rs
+++ b/compiler/rustc_parse/src/parser/path.rs
@@ -126,7 +126,7 @@ impl<'a> Parser<'a> {
     /// ```
     fn recover_colon_before_qpath_proj(&mut self) -> bool {
         if !self.check_noexpect(&TokenKind::Colon)
-            || self.look_ahead(1, |t| !t.is_ident() || t.is_reserved_ident())
+            || self.look_ahead(1, |t| !t.is_non_reserved_ident())
         {
             return false;
         }
@@ -260,7 +260,7 @@ impl<'a> Parser<'a> {
                 if self.may_recover()
                     && style == PathStyle::Expr // (!)
                     && self.token == token::Colon
-                    && self.look_ahead(1, |token| token.is_ident() && !token.is_reserved_ident())
+                    && self.look_ahead(1, |token| token.is_non_reserved_ident())
                 {
                     // Emit a special error message for `a::b:c` to help users
                     // otherwise, `a: c` might have meant to introduce a new binding
@@ -334,9 +334,7 @@ impl<'a> Parser<'a> {
                     self.expect_gt().map_err(|mut err| {
                         // Try to recover a `:` into a `::`
                         if self.token == token::Colon
-                            && self.look_ahead(1, |token| {
-                                token.is_ident() && !token.is_reserved_ident()
-                            })
+                            && self.look_ahead(1, |token| token.is_non_reserved_ident())
                         {
                             err.cancel();
                             err = self.dcx().create_err(PathSingleColon {
diff --git a/compiler/rustc_parse/src/parser/stmt.rs b/compiler/rustc_parse/src/parser/stmt.rs
index c37cb0881c3..2fa6520f2a4 100644
--- a/compiler/rustc_parse/src/parser/stmt.rs
+++ b/compiler/rustc_parse/src/parser/stmt.rs
@@ -798,7 +798,7 @@ impl<'a> Parser<'a> {
         }
         if self.prev_token.is_reserved_ident() && self.prev_token.is_ident_named(kw::Await) {
             // Likely `foo.await bar`
-        } else if !self.prev_token.is_reserved_ident() && self.prev_token.is_ident() {
+        } else if self.prev_token.is_non_reserved_ident() {
             // Likely `foo bar`
         } else if self.prev_token.kind == token::Question {
             // `foo? bar`
diff --git a/compiler/rustc_parse/src/parser/ty.rs b/compiler/rustc_parse/src/parser/ty.rs
index 620a34044d1..d874a071cee 100644
--- a/compiler/rustc_parse/src/parser/ty.rs
+++ b/compiler/rustc_parse/src/parser/ty.rs
@@ -885,6 +885,7 @@ impl<'a> Parser<'a> {
             || self.check(exp!(Tilde))
             || self.check_keyword(exp!(For))
             || self.check(exp!(OpenParen))
+            || self.check(exp!(OpenBracket))
             || self.check_keyword(exp!(Const))
             || self.check_keyword(exp!(Async))
             || self.check_keyword(exp!(Use))
@@ -982,12 +983,12 @@ impl<'a> Parser<'a> {
         Ok(())
     }
 
-    /// Parses the modifiers that may precede a trait in a bound, e.g. `?Trait` or `~const Trait`.
+    /// Parses the modifiers that may precede a trait in a bound, e.g. `?Trait` or `[const] Trait`.
     ///
     /// If no modifiers are present, this does not consume any tokens.
     ///
     /// ```ebnf
-    /// CONSTNESS = [["~"] "const"]
+    /// CONSTNESS = [["["] "const" ["]"]]
     /// ASYNCNESS = ["async"]
     /// POLARITY = ["?" | "!"]
     /// ```
@@ -995,18 +996,7 @@ impl<'a> Parser<'a> {
     /// See `parse_generic_ty_bound` for the complete grammar of trait bound modifiers.
     fn parse_trait_bound_modifiers(&mut self) -> PResult<'a, TraitBoundModifiers> {
         let modifier_lo = self.token.span;
-        let constness = if self.eat(exp!(Tilde)) {
-            let tilde = self.prev_token.span;
-            self.expect_keyword(exp!(Const))?;
-            let span = tilde.to(self.prev_token.span);
-            self.psess.gated_spans.gate(sym::const_trait_impl, span);
-            BoundConstness::Maybe(span)
-        } else if self.eat_keyword(exp!(Const)) {
-            self.psess.gated_spans.gate(sym::const_trait_impl, self.prev_token.span);
-            BoundConstness::Always(self.prev_token.span)
-        } else {
-            BoundConstness::Never
-        };
+        let constness = self.parse_bound_constness()?;
 
         let asyncness = if self.token_uninterpolated_span().at_least_rust_2018()
             && self.eat_keyword(exp!(Async))
@@ -1068,13 +1058,41 @@ impl<'a> Parser<'a> {
         Ok(TraitBoundModifiers { constness, asyncness, polarity })
     }
 
+    pub fn parse_bound_constness(&mut self) -> PResult<'a, BoundConstness> {
+        // FIXME(const_trait_impl): remove `~const` parser support once bootstrap has the new syntax
+        // in rustfmt
+        Ok(if self.eat(exp!(Tilde)) {
+            let tilde = self.prev_token.span;
+            self.expect_keyword(exp!(Const))?;
+            let span = tilde.to(self.prev_token.span);
+            self.psess.gated_spans.gate(sym::const_trait_impl, span);
+            BoundConstness::Maybe(span)
+        } else if self.check(exp!(OpenBracket))
+            && self.look_ahead(1, |t| t.is_keyword(kw::Const))
+            && self.look_ahead(2, |t| *t == token::CloseBracket)
+        {
+            let start = self.token.span;
+            self.bump();
+            self.expect_keyword(exp!(Const)).unwrap();
+            self.bump();
+            let span = start.to(self.prev_token.span);
+            self.psess.gated_spans.gate(sym::const_trait_impl, span);
+            BoundConstness::Maybe(span)
+        } else if self.eat_keyword(exp!(Const)) {
+            self.psess.gated_spans.gate(sym::const_trait_impl, self.prev_token.span);
+            BoundConstness::Always(self.prev_token.span)
+        } else {
+            BoundConstness::Never
+        })
+    }
+
     /// Parses a type bound according to:
     /// ```ebnf
     /// TY_BOUND = TY_BOUND_NOPAREN | (TY_BOUND_NOPAREN)
     /// TY_BOUND_NOPAREN = [for<GENERIC_PARAMS> CONSTNESS ASYNCNESS | POLARITY] SIMPLE_PATH
     /// ```
     ///
-    /// For example, this grammar accepts `for<'a: 'b> ~const ?m::Trait<'a>`.
+    /// For example, this grammar accepts `for<'a: 'b> [const] ?m::Trait<'a>`.
     fn parse_generic_ty_bound(
         &mut self,
         lo: Span,
@@ -1101,7 +1119,7 @@ impl<'a> Parser<'a> {
         }
 
         // Recover erroneous lifetime bound with modifiers or binder.
-        // e.g. `T: for<'a> 'a` or `T: ~const 'a`.
+        // e.g. `T: for<'a> 'a` or `T: [const] 'a`.
         if self.token.is_lifetime() {
             let _: ErrorGuaranteed = self.error_lt_bound_with_modifiers(modifiers, binder_span);
             return self.parse_generic_lt_bound(lo, has_parens);
diff --git a/compiler/rustc_parse/src/validate_attr.rs b/compiler/rustc_parse/src/validate_attr.rs
index 8e6442353c3..8b1d864c995 100644
--- a/compiler/rustc_parse/src/validate_attr.rs
+++ b/compiler/rustc_parse/src/validate_attr.rs
@@ -286,14 +286,23 @@ fn emit_malformed_attribute(
     if matches!(
         name,
         sym::inline
+            | sym::may_dangle
+            | sym::rustc_as_ptr
+            | sym::rustc_pub_transparent
+            | sym::rustc_const_stable_indirect
             | sym::rustc_force_inline
             | sym::rustc_confusables
+            | sym::rustc_skip_during_method_dispatch
             | sym::repr
             | sym::align
             | sym::deprecated
             | sym::optimize
             | sym::cold
+            | sym::naked
+            | sym::no_mangle
             | sym::must_use
+            | sym::track_caller
+            | sym::link_name
     ) {
         return;
     }