Don't make pattern nonterminals match statement nonterminals

Author: compiler-errors

compiler/rustc_ast/src/token.rs

@@ -447,6 +447,9 @@ impl Token {
     }
 
     /// Returns `true` if the token can appear at the start of an expression.
+    ///
+    /// **NB**: Take care when modifying this function, since it will change
+    /// the stable set of tokens that are allowed to match an expr nonterminal.
     pub fn can_begin_expr(&self) -> bool {
         match self.uninterpolate().kind {
             Ident(name, is_raw)              =>
@@ -475,24 +478,34 @@ impl Token {
 
     /// Returns `true` if the token can appear at the start of a pattern.
     ///
-    /// Shamelessly borrowed from `can_begin_expr`, only used for diagnostics right now.
-    pub fn can_begin_pattern(&self) -> bool {
-        match self.uninterpolate().kind {
-            Ident(name, is_raw)              =>
-                ident_can_begin_expr(name, self.span, is_raw), // value name or keyword
-            | OpenDelim(Delimiter::Bracket | Delimiter::Parenthesis)  // tuple or array
-            | Literal(..)                        // literal
-            | BinOp(Minus)                       // unary minus
-            | BinOp(And)                         // reference
-            | AndAnd                             // double reference
-            // DotDotDot is no longer supported
-            | DotDot | DotDotDot | DotDotEq      // ranges
-            | Lt | BinOp(Shl)                    // associated path
-            | ModSep                    => true, // global path
-            Interpolated(ref nt) => matches!(&nt.0, NtLiteral(..) |
-                NtPat(..)     |
-                NtBlock(..)   |
-                NtPath(..)),
+    /// **NB**: Take care when modifying this function, since it will change
+    /// the stable set of tokens that are allowed to match a pat nonterminal.
+    pub fn can_begin_pattern(&self, allow_leading_or: bool) -> bool {
+        match &self.uninterpolate().kind {
+            Ident(..) |                          // box, ref, mut, and other identifiers (can stricten)
+            OpenDelim(Delimiter::Parenthesis) |  // tuple pattern
+            OpenDelim(Delimiter::Bracket) |      // slice pattern
+            BinOp(And) |                  // reference
+            BinOp(Minus) |                // negative literal
+            AndAnd |                             // double reference
+            Literal(_) |                        // literal
+            DotDot |                             // range pattern (future compat)
+            DotDotDot |                          // range pattern (future compat)
+            ModSep |                             // path
+            Lt |                                 // path (UFCS constant)
+            BinOp(Shl) => true,           // path (double UFCS)
+            // leading vert `|` or-pattern
+            BinOp(Or) => allow_leading_or,
+            Interpolated(nt) =>
+                matches!(&nt.0,
+                    | NtBlock(..)
+                    | NtExpr(..)
+                    | NtLiteral(..)
+                    | NtMeta(..)
+                    | NtPat(..)
+                    | NtPath(..)
+                    | NtTy(..)
+                ),
             _ => false,
         }
     }

compiler/rustc_parse/src/parser/nonterminal.rs

@@ -66,26 +66,8 @@ impl<'a> Parser<'a> {
                 token::Interpolated(nt) => may_be_ident(&nt.0),
                 _ => false,
             },
-            NonterminalKind::PatParam { .. } | NonterminalKind::PatWithOr => {
-                match &token.kind {
-                token::Ident(..) |                          // box, ref, mut, and other identifiers (can stricten)
-                token::OpenDelim(Delimiter::Parenthesis) |  // tuple pattern
-                token::OpenDelim(Delimiter::Bracket) |      // slice pattern
-                token::BinOp(token::And) |                  // reference
-                token::BinOp(token::Minus) |                // negative literal
-                token::AndAnd |                             // double reference
-                token::Literal(_) |                        // literal
-                token::DotDot |                             // range pattern (future compat)
-                token::DotDotDot |                          // range pattern (future compat)
-                token::ModSep |                             // path
-                token::Lt |                                 // path (UFCS constant)
-                token::BinOp(token::Shl) => true,           // path (double UFCS)
-                // leading vert `|` or-pattern
-                token::BinOp(token::Or) => matches!(kind, NonterminalKind::PatWithOr),
-                token::Interpolated(nt) => may_be_ident(&nt.0),
-                _ => false,
-            }
-            }
+            NonterminalKind::PatParam { .. } => token.can_begin_pattern(false),
+            NonterminalKind::PatWithOr => token.can_begin_pattern(true),
             NonterminalKind::Lifetime => match &token.kind {
                 token::Lifetime(_) => true,
                 token::Interpolated(nt) => {

compiler/rustc_parse/src/parser/pat.rs

@@ -349,7 +349,8 @@ impl<'a> Parser<'a> {
 
         let mut lo = self.token.span;
 
-        if self.token.is_keyword(kw::Let) && self.look_ahead(1, |tok| tok.can_begin_pattern()) {
+        if self.token.is_keyword(kw::Let) && self.look_ahead(1, |tok| tok.can_begin_pattern(false))
+        {
             self.bump();
             self.dcx().emit_err(RemoveLet { span: lo });
             lo = self.token.span;

tests/ui/pattern/patterns-dont-match-nt-statement.rs

@@ -0,0 +1,19 @@
+// check-pass
+
+// Make sure that a `stmt` nonterminal does not eagerly match against
+// a `pat`, since this will always cause a parse error...
+
+macro_rules! m {
+    ($pat:pat) => {};
+    ($stmt:stmt) => {};
+}
+
+macro_rules! m2 {
+    ($stmt:stmt) => {
+        m! { $stmt }
+    };
+}
+
+m2! { let x = 1 }
+
+fn main() {}