Merge from rustc

5 files changed, 60 insertions, 25 deletions

diff --git a/compiler/rustc_parse/src/lexer/mod.rs b/compiler/rustc_parse/src/lexer/mod.rs
index f027843e6b4..8761c23625b 100644
--- a/compiler/rustc_parse/src/lexer/mod.rs
+++ b/compiler/rustc_parse/src/lexer/mod.rs
@@ -79,7 +79,7 @@ impl<'a> StringReader<'a> {
     /// preceded by whitespace.
     fn next_token(&mut self) -> (Token, bool) {
         let mut preceded_by_whitespace = false;
-
+        let mut swallow_next_invalid = 0;
         // Skip trivial (whitespace & comments) tokens
         loop {
             let token = self.cursor.advance_token();
@@ -232,19 +232,34 @@ impl<'a> StringReader<'a> {
                 rustc_lexer::TokenKind::Percent => token::BinOp(token::Percent),
 
                 rustc_lexer::TokenKind::Unknown | rustc_lexer::TokenKind::InvalidIdent => {
-                    let c = self.str_from(start).chars().next().unwrap();
+                    // Don't emit diagnostics for sequences of the same invalid token
+                    if swallow_next_invalid > 0 {
+                        swallow_next_invalid -= 1;
+                        continue;
+                    }
+                    let mut it = self.str_from_to_end(start).chars();
+                    let c = it.next().unwrap();
+                    let repeats = it.take_while(|c1| *c1 == c).count();
                     let mut err =
-                        self.struct_err_span_char(start, self.pos, "unknown start of token", c);
+                        self.struct_err_span_char(start, self.pos + Pos::from_usize(repeats * c.len_utf8()), "unknown start of token", c);
                     // FIXME: the lexer could be used to turn the ASCII version of unicode
                     // homoglyphs, instead of keeping a table in `check_for_substitution`into the
                     // token. Ideally, this should be inside `rustc_lexer`. However, we should
                     // first remove compound tokens like `<<` from `rustc_lexer`, and then add
                     // fancier error recovery to it, as there will be less overall work to do this
                     // way.
-                    let token = unicode_chars::check_for_substitution(self, start, c, &mut err);
+                    let token = unicode_chars::check_for_substitution(self, start, c, &mut err, repeats+1);
                     if c == '\x00' {
                         err.help("source files must contain UTF-8 encoded text, unexpected null bytes might occur when a different encoding is used");
                     }
+                    if repeats > 0 {
+                        if repeats == 1 {
+                            err.note(format!("character appears once more"));
+                        } else {
+                            err.note(format!("character appears {repeats} more times"));
+                        }
+                        swallow_next_invalid = repeats;
+                    }
                     err.emit();
                     if let Some(token) = token {
                         token
@@ -486,6 +501,11 @@ impl<'a> StringReader<'a> {
         &self.src[self.src_index(start)..self.src_index(end)]
     }
 
+    /// Slice of the source text spanning from `start` until the end
+    fn str_from_to_end(&self, start: BytePos) -> &str {
+        &self.src[self.src_index(start)..]
+    }
+
     fn report_raw_str_error(&self, start: BytePos, prefix_len: u32) -> ! {
         match rustc_lexer::validate_raw_str(self.str_from(start), prefix_len) {
             Err(RawStrError::InvalidStarter { bad_char }) => {
diff --git a/compiler/rustc_parse/src/lexer/unicode_chars.rs b/compiler/rustc_parse/src/lexer/unicode_chars.rs
index f1b50296e25..65479b341d7 100644
--- a/compiler/rustc_parse/src/lexer/unicode_chars.rs
+++ b/compiler/rustc_parse/src/lexer/unicode_chars.rs
@@ -337,10 +337,11 @@ pub(super) fn check_for_substitution<'a>(
     pos: BytePos,
     ch: char,
     err: &mut Diagnostic,
+    count: usize,
 ) -> Option<token::TokenKind> {
     let &(_u_char, u_name, ascii_char) = UNICODE_ARRAY.iter().find(|&&(c, _, _)| c == ch)?;
 
-    let span = Span::with_root_ctxt(pos, pos + Pos::from_usize(ch.len_utf8()));
+    let span = Span::with_root_ctxt(pos, pos + Pos::from_usize(ch.len_utf8() * count));
 
     let Some((_ascii_char, ascii_name, token)) = ASCII_ARRAY.iter().find(|&&(c, _, _)| c == ascii_char) else {
         let msg = format!("substitution character not found for '{}'", ch);
@@ -369,7 +370,12 @@ pub(super) fn check_for_substitution<'a>(
             "Unicode character '{}' ({}) looks like '{}' ({}), but it is not",
             ch, u_name, ascii_char, ascii_name
         );
-        err.span_suggestion(span, &msg, ascii_char, Applicability::MaybeIncorrect);
+        err.span_suggestion(
+            span,
+            &msg,
+            ascii_char.to_string().repeat(count),
+            Applicability::MaybeIncorrect,
+        );
     }
     token.clone()
 }
diff --git a/compiler/rustc_parse/src/parser/expr.rs b/compiler/rustc_parse/src/parser/expr.rs
index dd2b03988c3..d58afcd4c9f 100644
--- a/compiler/rustc_parse/src/parser/expr.rs
+++ b/compiler/rustc_parse/src/parser/expr.rs
@@ -83,7 +83,7 @@ macro_rules! maybe_whole_expr {
 pub(super) enum LhsExpr {
     NotYetParsed,
     AttributesParsed(AttrWrapper),
-    AlreadyParsed(P<Expr>, bool), // (expr, starts_statement)
+    AlreadyParsed { expr: P<Expr>, starts_statement: bool },
 }
 
 impl From<Option<AttrWrapper>> for LhsExpr {
@@ -97,11 +97,11 @@ impl From<Option<AttrWrapper>> for LhsExpr {
 }
 
 impl From<P<Expr>> for LhsExpr {
-    /// Converts the `expr: P<Expr>` into `LhsExpr::AlreadyParsed(expr)`.
+    /// Converts the `expr: P<Expr>` into `LhsExpr::AlreadyParsed { expr, starts_statement: false }`.
     ///
     /// This conversion does not allocate.
     fn from(expr: P<Expr>) -> Self {
-        LhsExpr::AlreadyParsed(expr, false)
+        LhsExpr::AlreadyParsed { expr, starts_statement: false }
     }
 }
 
@@ -174,7 +174,7 @@ impl<'a> Parser<'a> {
         lhs: LhsExpr,
     ) -> PResult<'a, P<Expr>> {
         let mut starts_stmt = false;
-        let mut lhs = if let LhsExpr::AlreadyParsed(expr, starts_statement) = lhs {
+        let mut lhs = if let LhsExpr::AlreadyParsed { expr, starts_statement } = lhs {
             starts_stmt = starts_statement;
             expr
         } else {
@@ -562,17 +562,23 @@ impl<'a> Parser<'a> {
 
         // Note: when adding new unary operators, don't forget to adjust TokenKind::can_begin_expr()
         match this.token.uninterpolate().kind {
-            token::Not => make_it!(this, attrs, |this, _| this.parse_unary_expr(lo, UnOp::Not)), // `!expr`
-            token::Tilde => make_it!(this, attrs, |this, _| this.recover_tilde_expr(lo)), // `~expr`
+            // `!expr`
+            token::Not => make_it!(this, attrs, |this, _| this.parse_unary_expr(lo, UnOp::Not)),
+            // `~expr`
+            token::Tilde => make_it!(this, attrs, |this, _| this.recover_tilde_expr(lo)),
+            // `-expr`
             token::BinOp(token::Minus) => {
                 make_it!(this, attrs, |this, _| this.parse_unary_expr(lo, UnOp::Neg))
-            } // `-expr`
+            }
+            // `*expr`
             token::BinOp(token::Star) => {
                 make_it!(this, attrs, |this, _| this.parse_unary_expr(lo, UnOp::Deref))
-            } // `*expr`
+            }
+            // `&expr` and `&&expr`
             token::BinOp(token::And) | token::AndAnd => {
                 make_it!(this, attrs, |this, _| this.parse_borrow_expr(lo))
             }
+            // `+lit`
             token::BinOp(token::Plus) if this.look_ahead(1, |tok| tok.is_numeric_lit()) => {
                 let mut err =
                     LeadingPlusNotSupported { span: lo, remove_plus: None, add_parentheses: None };
@@ -587,7 +593,7 @@ impl<'a> Parser<'a> {
 
                 this.bump();
                 this.parse_prefix_expr(None)
-            } // `+expr`
+            }
             // Recover from `++x`:
             token::BinOp(token::Plus)
                 if this.look_ahead(1, |t| *t == token::BinOp(token::Plus)) =>
@@ -624,7 +630,7 @@ impl<'a> Parser<'a> {
         Ok((span, self.mk_unary(op, expr)))
     }
 
-    // Recover on `!` suggesting for bitwise negation instead.
+    /// Recover on `~expr` in favor of `!expr`.
     fn recover_tilde_expr(&mut self, lo: Span) -> PResult<'a, (Span, ExprKind)> {
         self.sess.emit_err(TildeAsUnaryOperator(lo));
 
@@ -651,7 +657,6 @@ impl<'a> Parser<'a> {
 
     /// Recover on `not expr` in favor of `!expr`.
     fn recover_not_expr(&mut self, lo: Span) -> PResult<'a, (Span, ExprKind)> {
-        // Emit the error...
         let negated_token = self.look_ahead(1, |t| t.clone());
 
         let sub_diag = if negated_token.is_numeric_lit() {
@@ -672,7 +677,6 @@ impl<'a> Parser<'a> {
             ),
         });
 
-        // ...and recover!
         self.parse_unary_expr(lo, UnOp::Not)
     }
 
@@ -1471,9 +1475,8 @@ impl<'a> Parser<'a> {
             } else if self.eat(&token::Comma) {
                 // Vector with two or more elements.
                 let sep = SeqSep::trailing_allowed(token::Comma);
-                let (remaining_exprs, _) = self.parse_seq_to_end(close, sep, |p| p.parse_expr())?;
-                let mut exprs = vec![first_expr];
-                exprs.extend(remaining_exprs);
+                let (mut exprs, _) = self.parse_seq_to_end(close, sep, |p| p.parse_expr())?;
+                exprs.insert(0, first_expr);
                 ExprKind::Array(exprs)
             } else {
                 // Vector with one element
@@ -1593,7 +1596,7 @@ impl<'a> Parser<'a> {
                     vis.0
                 };
 
-                // Suggestion involves adding a (as of time of writing this, unstable) labeled block.
+                // Suggestion involves adding a labeled block.
                 //
                 // If there are no breaks that may use this label, suggest removing the label and
                 // recover to the unmodified expression.
diff --git a/compiler/rustc_parse/src/parser/pat.rs b/compiler/rustc_parse/src/parser/pat.rs
index 0b057f2f577..e73a17ced7d 100644
--- a/compiler/rustc_parse/src/parser/pat.rs
+++ b/compiler/rustc_parse/src/parser/pat.rs
@@ -469,7 +469,7 @@ impl<'a> Parser<'a> {
     /// Try to recover the more general form `intersect ::= $pat_lhs @ $pat_rhs`.
     ///
     /// Allowed binding patterns generated by `binding ::= ref? mut? $ident @ $pat_rhs`
-    /// should already have been parsed by now  at this point,
+    /// should already have been parsed by now at this point,
     /// if the next token is `@` then we can try to parse the more general form.
     ///
     /// Consult `parse_pat_ident` for the `binding` grammar.
diff --git a/compiler/rustc_parse/src/parser/stmt.rs b/compiler/rustc_parse/src/parser/stmt.rs
index 1e5c2834960..4ff9927aab5 100644
--- a/compiler/rustc_parse/src/parser/stmt.rs
+++ b/compiler/rustc_parse/src/parser/stmt.rs
@@ -164,7 +164,10 @@ impl<'a> Parser<'a> {
             // Perform this outside of the `collect_tokens_trailing_token` closure,
             // since our outer attributes do not apply to this part of the expression
             let expr = self.with_res(Restrictions::STMT_EXPR, |this| {
-                this.parse_assoc_expr_with(0, LhsExpr::AlreadyParsed(expr, true))
+                this.parse_assoc_expr_with(
+                    0,
+                    LhsExpr::AlreadyParsed { expr, starts_statement: true },
+                )
             })?;
             Ok(self.mk_stmt(lo.to(self.prev_token.span), StmtKind::Expr(expr)))
         } else {
@@ -198,7 +201,10 @@ impl<'a> Parser<'a> {
             let e = self.mk_expr(lo.to(hi), ExprKind::MacCall(mac));
             let e = self.maybe_recover_from_bad_qpath(e)?;
             let e = self.parse_dot_or_call_expr_with(e, lo, attrs)?;
-            let e = self.parse_assoc_expr_with(0, LhsExpr::AlreadyParsed(e, false))?;
+            let e = self.parse_assoc_expr_with(
+                0,
+                LhsExpr::AlreadyParsed { expr: e, starts_statement: false },
+            )?;
             StmtKind::Expr(e)
         };
         Ok(self.mk_stmt(lo.to(hi), kind))