Auto merge of #132629 - nnethercote:124141-preliminaries, r=petrochenkov

#124141 preliminaries

Preliminary changes required to start removing `Nonterminal` (#124141).

r? `@petrochenkov`

compiler/rustc_ast/src/attr/mod.rs

@@ -457,7 +457,7 @@ impl MetaItemKind {
         tokens: &mut impl Iterator<Item = &'a TokenTree>,
     ) -> Option<MetaItemKind> {
         match tokens.next() {
-            Some(TokenTree::Delimited(.., Delimiter::Invisible, inner_tokens)) => {
+            Some(TokenTree::Delimited(.., Delimiter::Invisible(_), inner_tokens)) => {
                 MetaItemKind::name_value_from_tokens(&mut inner_tokens.trees())
             }
             Some(TokenTree::Token(token, _)) => {
@@ -605,7 +605,7 @@ impl MetaItemInner {
                 tokens.next();
                 return Some(MetaItemInner::Lit(lit));
             }
-            Some(TokenTree::Delimited(.., Delimiter::Invisible, inner_tokens)) => {
+            Some(TokenTree::Delimited(.., Delimiter::Invisible(_), inner_tokens)) => {
                 tokens.next();
                 return MetaItemInner::from_tokens(&mut inner_tokens.trees().peekable());
             }

compiler/rustc_ast/src/token.rs

@@ -42,11 +42,86 @@ pub enum BinOpToken {
     Shr,
 }
 
+// This type must not implement `Hash` due to the unusual `PartialEq` impl below.
+#[derive(Copy, Clone, Debug, Encodable, Decodable, HashStable_Generic)]
+pub enum InvisibleOrigin {
+    // From the expansion of a metavariable in a declarative macro.
+    MetaVar(MetaVarKind),
+
+    // Converted from `proc_macro::Delimiter` in
+    // `proc_macro::Delimiter::to_internal`, i.e. returned by a proc macro.
+    ProcMacro,
+
+    // Converted from `TokenKind::Interpolated` in
+    // `TokenStream::flatten_token`. Treated similarly to `ProcMacro`.
+    FlattenToken,
+}
+
+impl PartialEq for InvisibleOrigin {
+    #[inline]
+    fn eq(&self, _other: &InvisibleOrigin) -> bool {
+        // When we had AST-based nonterminals we couldn't compare them, and the
+        // old `Nonterminal` type had an `eq` that always returned false,
+        // resulting in this restriction:
+        // https://doc.rust-lang.org/nightly/reference/macros-by-example.html#forwarding-a-matched-fragment
+        // This `eq` emulates that behaviour. We could consider lifting this
+        // restriction now but there are still cases involving invisible
+        // delimiters that make it harder than it first appears.
+        false
+    }
+}
+
+/// Annoyingly similar to `NonterminalKind`, but the slight differences are important.
+#[derive(Debug, Copy, Clone, PartialEq, Eq, Encodable, Decodable, Hash, HashStable_Generic)]
+pub enum MetaVarKind {
+    Item,
+    Block,
+    Stmt,
+    Pat(NtPatKind),
+    Expr {
+        kind: NtExprKind,
+        // This field is needed for `Token::can_begin_literal_maybe_minus`.
+        can_begin_literal_maybe_minus: bool,
+        // This field is needed for `Token::can_begin_string_literal`.
+        can_begin_string_literal: bool,
+    },
+    Ty,
+    Ident,
+    Lifetime,
+    Literal,
+    Meta,
+    Path,
+    Vis,
+    TT,
+}
+
+impl fmt::Display for MetaVarKind {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        let sym = match self {
+            MetaVarKind::Item => sym::item,
+            MetaVarKind::Block => sym::block,
+            MetaVarKind::Stmt => sym::stmt,
+            MetaVarKind::Pat(PatParam { inferred: true } | PatWithOr) => sym::pat,
+            MetaVarKind::Pat(PatParam { inferred: false }) => sym::pat_param,
+            MetaVarKind::Expr { kind: Expr2021 { inferred: true } | Expr, .. } => sym::expr,
+            MetaVarKind::Expr { kind: Expr2021 { inferred: false }, .. } => sym::expr_2021,
+            MetaVarKind::Ty => sym::ty,
+            MetaVarKind::Ident => sym::ident,
+            MetaVarKind::Lifetime => sym::lifetime,
+            MetaVarKind::Literal => sym::literal,
+            MetaVarKind::Meta => sym::meta,
+            MetaVarKind::Path => sym::path,
+            MetaVarKind::Vis => sym::vis,
+            MetaVarKind::TT => sym::tt,
+        };
+        write!(f, "{sym}")
+    }
+}
+
 /// Describes how a sequence of token trees is delimited.
 /// Cannot use `proc_macro::Delimiter` directly because this
 /// structure should implement some additional traits.
-#[derive(Copy, Clone, Debug, PartialEq, Eq)]
-#[derive(Encodable, Decodable, Hash, HashStable_Generic)]
+#[derive(Copy, Clone, Debug, PartialEq, Encodable, Decodable, HashStable_Generic)]
 pub enum Delimiter {
     /// `( ... )`
     Parenthesis,
@@ -59,7 +134,34 @@ pub enum Delimiter {
     /// "macro variable" `$var`. It is important to preserve operator priorities in cases like
     /// `$var * 3` where `$var` is `1 + 2`.
     /// Invisible delimiters might not survive roundtrip of a token stream through a string.
-    Invisible,
+    Invisible(InvisibleOrigin),
+}
+
+impl Delimiter {
+    // Should the parser skip these delimiters? Only happens for certain kinds
+    // of invisible delimiters. Ideally this function will eventually disappear
+    // and no invisible delimiters will be skipped.
+    #[inline]
+    pub fn skip(&self) -> bool {
+        match self {
+            Delimiter::Parenthesis | Delimiter::Bracket | Delimiter::Brace => false,
+            Delimiter::Invisible(InvisibleOrigin::MetaVar(_)) => false,
+            Delimiter::Invisible(InvisibleOrigin::FlattenToken | InvisibleOrigin::ProcMacro) => {
+                true
+            }
+        }
+    }
+
+    // This exists because `InvisibleOrigin`s should be compared. It is only used for assertions.
+    pub fn eq_ignoring_invisible_origin(&self, other: &Delimiter) -> bool {
+        match (self, other) {
+            (Delimiter::Parenthesis, Delimiter::Parenthesis) => true,
+            (Delimiter::Brace, Delimiter::Brace) => true,
+            (Delimiter::Bracket, Delimiter::Bracket) => true,
+            (Delimiter::Invisible(_), Delimiter::Invisible(_)) => true,
+            _ => false,
+        }
+    }
 }
 
 // Note that the suffix is *not* considered when deciding the `LitKind` in this
@@ -496,10 +598,11 @@ impl Token {
     /// **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 {
+        use Delimiter::*;
         match self.uninterpolate().kind {
             Ident(name, is_raw)              =>
                 ident_can_begin_expr(name, self.span, is_raw), // value name or keyword
-            OpenDelim(..)                     | // tuple, array or block
+            OpenDelim(Parenthesis | Brace | Bracket) | // tuple, array or block
             Literal(..)                       | // literal
             Not                               | // operator not
             BinOp(Minus)                      | // unary minus
@@ -510,7 +613,7 @@ impl Token {
             // DotDotDot is no longer supported, but we need some way to display the error
             DotDot | DotDotDot | DotDotEq     | // range notation
             Lt | BinOp(Shl)                   | // associated path
-            PathSep                            | // global path
+            PathSep                           | // global path
             Lifetime(..)                      | // labeled loop
             Pound                             => true, // expression attributes
             Interpolated(ref nt) =>
@@ -520,6 +623,12 @@ impl Token {
                     NtLiteral(..) |
                     NtPath(..)
                 ),
+            OpenDelim(Delimiter::Invisible(InvisibleOrigin::MetaVar(
+                MetaVarKind::Block |
+                MetaVarKind::Expr { .. } |
+                MetaVarKind::Literal |
+                MetaVarKind::Path
+            ))) => true,
             _ => false,
         }
     }
@@ -553,6 +662,14 @@ impl Token {
                     | NtPath(..)
                     | NtTy(..)
                 ),
+            OpenDelim(Delimiter::Invisible(InvisibleOrigin::MetaVar(
+                MetaVarKind::Expr { .. } |
+                MetaVarKind::Literal |
+                MetaVarKind::Meta |
+                MetaVarKind::Pat(_) |
+                MetaVarKind::Path |
+                MetaVarKind::Ty
+            ))) => true,
             _ => false,
         }
     }
@@ -573,6 +690,10 @@ impl Token {
             Lt | BinOp(Shl)             | // associated path
             PathSep                      => true, // global path
             Interpolated(ref nt) => matches!(&**nt, NtTy(..) | NtPath(..)),
+            OpenDelim(Delimiter::Invisible(InvisibleOrigin::MetaVar(
+                MetaVarKind::Ty |
+                MetaVarKind::Path
+            ))) => true,
             // For anonymous structs or unions, which only appear in specific positions
             // (type of struct fields or union fields), we don't consider them as regular types
             _ => false,
@@ -585,6 +706,9 @@ impl Token {
             OpenDelim(Delimiter::Brace) | Literal(..) | BinOp(Minus) => true,
             Ident(name, IdentIsRaw::No) if name.is_bool_lit() => true,
             Interpolated(ref nt) => matches!(&**nt, NtExpr(..) | NtBlock(..) | NtLiteral(..)),
+            OpenDelim(Delimiter::Invisible(InvisibleOrigin::MetaVar(
+                MetaVarKind::Expr { .. } | MetaVarKind::Block | MetaVarKind::Literal,
+            ))) => true,
             _ => false,
         }
     }
@@ -641,6 +765,13 @@ impl Token {
                 },
                 _ => false,
             },
+            OpenDelim(Delimiter::Invisible(InvisibleOrigin::MetaVar(mv_kind))) => match mv_kind {
+                MetaVarKind::Literal => true,
+                MetaVarKind::Expr { can_begin_literal_maybe_minus, .. } => {
+                    can_begin_literal_maybe_minus
+                }
+                _ => false,
+            },
             _ => false,
         }
     }
@@ -656,6 +787,11 @@ impl Token {
                 },
                 _ => false,
             },
+            OpenDelim(Delimiter::Invisible(InvisibleOrigin::MetaVar(mv_kind))) => match mv_kind {
+                MetaVarKind::Literal => true,
+                MetaVarKind::Expr { can_begin_string_literal, .. } => can_begin_string_literal,
+                _ => false,
+            },
             _ => false,
         }
     }
@@ -896,7 +1032,7 @@ impl PartialEq<TokenKind> for Token {
     }
 }
 
-#[derive(Debug, Copy, Clone, PartialEq, Eq, Encodable, Decodable)]
+#[derive(Debug, Copy, Clone, PartialEq, Eq, Encodable, Decodable, Hash, HashStable_Generic)]
 pub enum NtPatKind {
     // Matches or-patterns. Was written using `pat` in edition 2021 or later.
     PatWithOr,
@@ -906,7 +1042,7 @@ pub enum NtPatKind {
     PatParam { inferred: bool },
 }
 
-#[derive(Debug, Copy, Clone, PartialEq, Eq, Encodable, Decodable)]
+#[derive(Debug, Copy, Clone, PartialEq, Eq, Encodable, Decodable, Hash, HashStable_Generic)]
 pub enum NtExprKind {
     // Matches expressions using the post-edition 2024. Was written using
     // `expr` in edition 2024 or later.
@@ -933,7 +1069,7 @@ pub enum Nonterminal {
     NtVis(P<ast::Visibility>),
 }
 
-#[derive(Debug, Copy, Clone, PartialEq, Encodable, Decodable)]
+#[derive(Debug, Copy, Clone, PartialEq, Eq, Encodable, Decodable, Hash, HashStable_Generic)]
 pub enum NonterminalKind {
     Item,
     Block,

compiler/rustc_ast/src/tokenstream.rs

@@ -24,7 +24,7 @@ use rustc_span::{DUMMY_SP, Span, SpanDecoder, SpanEncoder, Symbol, sym};
 
 use crate::ast::{AttrStyle, StmtKind};
 use crate::ast_traits::{HasAttrs, HasTokens};
-use crate::token::{self, Delimiter, Nonterminal, Token, TokenKind};
+use crate::token::{self, Delimiter, InvisibleOrigin, Nonterminal, Token, TokenKind};
 use crate::{AttrVec, Attribute};
 
 /// Part of a `TokenStream`.
@@ -484,13 +484,13 @@ impl TokenStream {
             token::NtLifetime(ident, is_raw) => TokenTree::Delimited(
                 DelimSpan::from_single(token.span),
                 DelimSpacing::new(Spacing::JointHidden, spacing),
-                Delimiter::Invisible,
+                Delimiter::Invisible(InvisibleOrigin::FlattenToken),
                 TokenStream::token_alone(token::Lifetime(ident.name, is_raw), ident.span),
             ),
             token::Interpolated(ref nt) => TokenTree::Delimited(
                 DelimSpan::from_single(token.span),
                 DelimSpacing::new(Spacing::JointHidden, spacing),
-                Delimiter::Invisible,
+                Delimiter::Invisible(InvisibleOrigin::FlattenToken),
                 TokenStream::from_nonterminal_ast(&nt).flattened(),
             ),
             _ => TokenTree::Token(token.clone(), spacing),

compiler/rustc_ast_pretty/src/pprust/state.rs

@@ -942,9 +942,8 @@ pub trait PrintState<'a>: std::ops::Deref<Target = pp::Printer> + std::ops::Dere
             token::CloseDelim(Delimiter::Bracket) => "]".into(),
             token::OpenDelim(Delimiter::Brace) => "{".into(),
             token::CloseDelim(Delimiter::Brace) => "}".into(),
-            token::OpenDelim(Delimiter::Invisible) | token::CloseDelim(Delimiter::Invisible) => {
-                "".into()
-            }
+            token::OpenDelim(Delimiter::Invisible(_))
+            | token::CloseDelim(Delimiter::Invisible(_)) => "".into(),
             token::Pound => "#".into(),
             token::Dollar => "$".into(),
             token::Question => "?".into(),

compiler/rustc_expand/src/mbe/diagnostics.rs

@@ -1,6 +1,6 @@
 use std::borrow::Cow;
 
-use rustc_ast::token::{self, Token, TokenKind};
+use rustc_ast::token::{self, Delimiter, Token, TokenKind};
 use rustc_ast::tokenstream::TokenStream;
 use rustc_errors::{Applicability, Diag, DiagCtxtHandle, DiagMessage};
 use rustc_macros::Subdiagnostic;
@@ -68,7 +68,9 @@ pub(super) fn failed_to_match_macro(
 
     if let MatcherLoc::Token { token: expected_token } = &remaining_matcher
         && (matches!(expected_token.kind, TokenKind::Interpolated(_))
-            || matches!(token.kind, TokenKind::Interpolated(_)))
+            || matches!(token.kind, TokenKind::Interpolated(_))
+            || matches!(expected_token.kind, TokenKind::OpenDelim(Delimiter::Invisible(_)))
+            || matches!(token.kind, TokenKind::OpenDelim(Delimiter::Invisible(_))))
     {
         err.note("captured metavariables except for `:tt`, `:ident` and `:lifetime` cannot be compared to other tokens");
         err.note("see <https://doc.rust-lang.org/nightly/reference/macros-by-example.html#forwarding-a-matched-fragment> for more information");

compiler/rustc_expand/src/mbe/macro_rules.rs

@@ -693,7 +693,7 @@ fn has_compile_error_macro(rhs: &mbe::TokenTree) -> bool {
                     && let mbe::TokenTree::Token(bang) = bang
                     && let TokenKind::Not = bang.kind
                     && let mbe::TokenTree::Delimited(.., del) = args
-                    && del.delim != Delimiter::Invisible
+                    && !del.delim.skip()
                 {
                     true
                 } else {

compiler/rustc_expand/src/mbe/quoted.rs

@@ -165,11 +165,12 @@ fn parse_tree<'a>(
             // during parsing.
             let mut next = outer_trees.next();
             let mut trees: Box<dyn Iterator<Item = &tokenstream::TokenTree>>;
-            if let Some(tokenstream::TokenTree::Delimited(.., Delimiter::Invisible, tts)) = next {
-                trees = Box::new(tts.trees());
-                next = trees.next();
-            } else {
-                trees = Box::new(outer_trees);
+            match next {
+                Some(tokenstream::TokenTree::Delimited(.., delim, tts)) if delim.skip() => {
+                    trees = Box::new(tts.trees());
+                    next = trees.next();
+                }
+                _ => trees = Box::new(outer_trees),
             }
 
             match next {

compiler/rustc_expand/src/proc_macro_server.rs

@@ -38,7 +38,7 @@ impl FromInternal<token::Delimiter> for Delimiter {
             token::Delimiter::Parenthesis => Delimiter::Parenthesis,
             token::Delimiter::Brace => Delimiter::Brace,
             token::Delimiter::Bracket => Delimiter::Bracket,
-            token::Delimiter::Invisible => Delimiter::None,
+            token::Delimiter::Invisible(_) => Delimiter::None,
         }
     }
 }
@@ -49,7 +49,7 @@ impl ToInternal<token::Delimiter> for Delimiter {
             Delimiter::Parenthesis => token::Delimiter::Parenthesis,
             Delimiter::Brace => token::Delimiter::Brace,
             Delimiter::Bracket => token::Delimiter::Bracket,
-            Delimiter::None => token::Delimiter::Invisible,
+            Delimiter::None => token::Delimiter::Invisible(token::InvisibleOrigin::ProcMacro),
         }
     }
 }

compiler/rustc_parse/messages.ftl

@@ -216,6 +216,9 @@ parse_expected_identifier_found_doc_comment = expected identifier, found doc com
 parse_expected_identifier_found_doc_comment_str = expected identifier, found doc comment `{$token}`
 parse_expected_identifier_found_keyword = expected identifier, found keyword
 parse_expected_identifier_found_keyword_str = expected identifier, found keyword `{$token}`
+parse_expected_identifier_found_metavar = expected identifier, found metavariable
+# This one deliberately doesn't print a token.
+parse_expected_identifier_found_metavar_str = expected identifier, found metavariable
 parse_expected_identifier_found_reserved_identifier = expected identifier, found reserved identifier
 parse_expected_identifier_found_reserved_identifier_str = expected identifier, found reserved identifier `{$token}`
 parse_expected_identifier_found_reserved_keyword = expected identifier, found reserved keyword
@@ -227,6 +230,8 @@ parse_expected_mut_or_const_in_raw_pointer_type = expected `mut` or `const` keyw
 
 parse_expected_semi_found_doc_comment_str = expected `;`, found doc comment `{$token}`
 parse_expected_semi_found_keyword_str = expected `;`, found keyword `{$token}`
+# This one deliberately doesn't print a token.
+parse_expected_semi_found_metavar_str = expected `;`, found metavariable
 parse_expected_semi_found_reserved_identifier_str = expected `;`, found reserved identifier `{$token}`
 parse_expected_semi_found_reserved_keyword_str = expected `;`, found reserved keyword `{$token}`
 parse_expected_semi_found_str = expected `;`, found `{$token}`
@@ -864,6 +869,8 @@ parse_unexpected_token_after_not_logical = use `!` to perform logical negation
 parse_unexpected_token_after_struct_name = expected `where`, `{"{"}`, `(`, or `;` after struct name
 parse_unexpected_token_after_struct_name_found_doc_comment = expected `where`, `{"{"}`, `(`, or `;` after struct name, found doc comment `{$token}`
 parse_unexpected_token_after_struct_name_found_keyword = expected `where`, `{"{"}`, `(`, or `;` after struct name, found keyword `{$token}`
+# This one deliberately doesn't print a token.
+parse_unexpected_token_after_struct_name_found_metavar = expected `where`, `{"{"}`, `(`, or `;` after struct name, found metavar
 parse_unexpected_token_after_struct_name_found_other = expected `where`, `{"{"}`, `(`, or `;` after struct name, found `{$token}`
 
 parse_unexpected_token_after_struct_name_found_reserved_identifier = expected `where`, `{"{"}`, `(`, or `;` after struct name, found reserved identifier `{$token}`