Properly handle attributes on statements

We now collect tokens for the underlying node wrapped by `StmtKind`
instead of storing tokens directly in `Stmt`.

`LazyTokenStream` now supports capturing a trailing semicolon after it
is initially constructed. This allows us to avoid refactoring statement
parsing to wrap the parsing of the semicolon in `parse_tokens`.

Attributes on item statements
(e.g. `fn foo() { #[bar] struct MyStruct; }`) are now treated as
item attributes, not statement attributes, which is consistent with how
we handle attributes on other kinds of statements. The feature-gating
code is adjusted so that proc-macro attributes are still allowed on item
statements on stable.

Two built-in macros (`#[global_allocator]` and `#[test]`) needed to be
adjusted to support being passed `Annotatable::Stmt`.

compiler/rustc_ast/src/ast.rs

@@ -901,29 +901,65 @@ pub struct Stmt {
     pub id: NodeId,
     pub kind: StmtKind,
     pub span: Span,
-    pub tokens: Option<LazyTokenStream>,
 }
 
 impl Stmt {
+    pub fn tokens(&self) -> Option<&LazyTokenStream> {
+        match self.kind {
+            StmtKind::Local(ref local) => local.tokens.as_ref(),
+            StmtKind::Item(ref item) => item.tokens.as_ref(),
+            StmtKind::Expr(ref expr) | StmtKind::Semi(ref expr) => expr.tokens.as_ref(),
+            StmtKind::Empty => None,
+            StmtKind::MacCall(ref mac) => mac.tokens.as_ref(),
+        }
+    }
+
+    pub fn tokens_mut(&mut self) -> Option<&mut LazyTokenStream> {
+        match self.kind {
+            StmtKind::Local(ref mut local) => local.tokens.as_mut(),
+            StmtKind::Item(ref mut item) => item.tokens.as_mut(),
+            StmtKind::Expr(ref mut expr) | StmtKind::Semi(ref mut expr) => expr.tokens.as_mut(),
+            StmtKind::Empty => None,
+            StmtKind::MacCall(ref mut mac) => mac.tokens.as_mut(),
+        }
+    }
+
+    pub fn set_tokens(&mut self, tokens: Option<LazyTokenStream>) {
+        match self.kind {
+            StmtKind::Local(ref mut local) => local.tokens = tokens,
+            StmtKind::Item(ref mut item) => item.tokens = tokens,
+            StmtKind::Expr(ref mut expr) | StmtKind::Semi(ref mut expr) => expr.tokens = tokens,
+            StmtKind::Empty => {}
+            StmtKind::MacCall(ref mut mac) => mac.tokens = tokens,
+        }
+    }
+
     pub fn has_trailing_semicolon(&self) -> bool {
         match &self.kind {
             StmtKind::Semi(_) => true,
             StmtKind::MacCall(mac) => matches!(mac.style, MacStmtStyle::Semicolon),
             _ => false,
         }
     }
+
+    /// Converts a parsed `Stmt` to a `Stmt` with
+    /// a trailing semicolon.
+    ///
+    /// This only modifies the parsed AST struct, not the attached
+    /// `LazyTokenStream`. The parser is responsible for calling
+    /// `CreateTokenStream::add_trailing_semi` when there is actually
+    /// a semicolon in the tokenstream.
     pub fn add_trailing_semicolon(mut self) -> Self {
         self.kind = match self.kind {
             StmtKind::Expr(expr) => StmtKind::Semi(expr),
             StmtKind::MacCall(mac) => {
-                StmtKind::MacCall(mac.map(|MacCallStmt { mac, style: _, attrs }| MacCallStmt {
-                    mac,
-                    style: MacStmtStyle::Semicolon,
-                    attrs,
+                StmtKind::MacCall(mac.map(|MacCallStmt { mac, style: _, attrs, tokens }| {
+                    MacCallStmt { mac, style: MacStmtStyle::Semicolon, attrs, tokens }
                 }))
             }
             kind => kind,
         };
+
         self
     }
 
@@ -963,6 +999,7 @@ pub struct MacCallStmt {
     pub mac: MacCall,
     pub style: MacStmtStyle,
     pub attrs: AttrVec,
+    pub tokens: Option<LazyTokenStream>,
 }
 
 #[derive(Clone, Copy, PartialEq, Encodable, Decodable, Debug)]
@@ -988,6 +1025,7 @@ pub struct Local {
     pub init: Option<P<Expr>>,
     pub span: Span,
     pub attrs: AttrVec,
+    pub tokens: Option<LazyTokenStream>,
 }
 
 /// An arm of a 'match'.

compiler/rustc_ast/src/mut_visit.rs

@@ -576,13 +576,14 @@ pub fn noop_visit_parenthesized_parameter_data<T: MutVisitor>(
 }
 
 pub fn noop_visit_local<T: MutVisitor>(local: &mut P<Local>, vis: &mut T) {
-    let Local { id, pat, ty, init, span, attrs } = local.deref_mut();
+    let Local { id, pat, ty, init, span, attrs, tokens } = local.deref_mut();
     vis.visit_id(id);
     vis.visit_pat(pat);
     visit_opt(ty, |ty| vis.visit_ty(ty));
     visit_opt(init, |init| vis.visit_expr(init));
     vis.visit_span(span);
     visit_thin_attrs(attrs, vis);
+    visit_lazy_tts(tokens, vis);
 }
 
 pub fn noop_visit_attribute<T: MutVisitor>(attr: &mut Attribute, vis: &mut T) {
@@ -1325,16 +1326,12 @@ pub fn noop_filter_map_expr<T: MutVisitor>(mut e: P<Expr>, vis: &mut T) -> Optio
 }
 
 pub fn noop_flat_map_stmt<T: MutVisitor>(
-    Stmt { kind, mut span, mut id, mut tokens }: Stmt,
+    Stmt { kind, mut span, mut id }: Stmt,
     vis: &mut T,
 ) -> SmallVec<[Stmt; 1]> {
     vis.visit_id(&mut id);
     vis.visit_span(&mut span);
-    visit_lazy_tts(&mut tokens, vis);
-    noop_flat_map_stmt_kind(kind, vis)
-        .into_iter()
-        .map(|kind| Stmt { id, kind, span, tokens: tokens.clone() })
-        .collect()
+    noop_flat_map_stmt_kind(kind, vis).into_iter().map(|kind| Stmt { id, kind, span }).collect()
 }
 
 pub fn noop_flat_map_stmt_kind<T: MutVisitor>(
@@ -1351,9 +1348,10 @@ pub fn noop_flat_map_stmt_kind<T: MutVisitor>(
         StmtKind::Semi(expr) => vis.filter_map_expr(expr).into_iter().map(StmtKind::Semi).collect(),
         StmtKind::Empty => smallvec![StmtKind::Empty],
         StmtKind::MacCall(mut mac) => {
-            let MacCallStmt { mac: mac_, style: _, attrs } = mac.deref_mut();
+            let MacCallStmt { mac: mac_, style: _, attrs, tokens } = mac.deref_mut();
             vis.visit_mac_call(mac_);
             visit_thin_attrs(attrs, vis);
+            visit_lazy_tts(tokens, vis);
             smallvec![StmtKind::MacCall(mac)]
         }
     }

compiler/rustc_ast/src/tokenstream.rs

@@ -121,10 +121,14 @@ where
 }
 
 pub trait CreateTokenStream: sync::Send + sync::Sync {
+    fn add_trailing_semi(&self) -> Box<dyn CreateTokenStream>;
     fn create_token_stream(&self) -> TokenStream;
 }
 
 impl CreateTokenStream for TokenStream {
+    fn add_trailing_semi(&self) -> Box<dyn CreateTokenStream> {
+        panic!("Cannot call `add_trailing_semi` on a `TokenStream`!");
+    }
     fn create_token_stream(&self) -> TokenStream {
         self.clone()
     }
@@ -141,6 +145,13 @@ impl LazyTokenStream {
         LazyTokenStream(Lrc::new(Box::new(inner)))
     }
 
+    /// Extends the captured stream by one token,
+    /// which must be a trailing semicolon. This
+    /// affects the `TokenStream` created by `make_tokenstream`.
+    pub fn add_trailing_semi(&self) -> LazyTokenStream {
+        LazyTokenStream(Lrc::new(self.0.add_trailing_semi()))
+    }
+
     pub fn create_token_stream(&self) -> TokenStream {
         self.0.create_token_stream()
     }

compiler/rustc_ast/src/visit.rs

@@ -686,7 +686,7 @@ pub fn walk_stmt<'a, V: Visitor<'a>>(visitor: &mut V, statement: &'a Stmt) {
         StmtKind::Expr(ref expr) | StmtKind::Semi(ref expr) => visitor.visit_expr(expr),
         StmtKind::Empty => {}
         StmtKind::MacCall(ref mac) => {
-            let MacCallStmt { ref mac, style: _, ref attrs } = **mac;
+            let MacCallStmt { ref mac, style: _, ref attrs, tokens: _ } = **mac;
             visitor.visit_mac_call(mac);
             for attr in attrs.iter() {
                 visitor.visit_attribute(attr);

compiler/rustc_builtin_macros/src/deriving/debug.rs

@@ -132,6 +132,7 @@ fn stmt_let_underscore(cx: &mut ExtCtxt<'_>, sp: Span, expr: P<ast::Expr>) -> as
         id: ast::DUMMY_NODE_ID,
         span: sp,
         attrs: ast::AttrVec::new(),
+        tokens: None,
     });
-    ast::Stmt { id: ast::DUMMY_NODE_ID, kind: ast::StmtKind::Local(local), span: sp, tokens: None }
+    ast::Stmt { id: ast::DUMMY_NODE_ID, kind: ast::StmtKind::Local(local), span: sp }
 }

compiler/rustc_builtin_macros/src/deriving/mod.rs

@@ -64,7 +64,6 @@ impl MultiItemModifier for BuiltinDerive {
                             id: ast::DUMMY_NODE_ID,
                             kind: ast::StmtKind::Item(a.expect_item()),
                             span,
-                            tokens: None,
                         })));
                     });
                 } else {

compiler/rustc_expand/src/base.rs

@@ -374,7 +374,6 @@ macro_rules! make_stmts_default {
                 id: ast::DUMMY_NODE_ID,
                 span: e.span,
                 kind: ast::StmtKind::Expr(e),
-                tokens: None
             }]
         })
     };
@@ -617,7 +616,6 @@ impl MacResult for DummyResult {
             id: ast::DUMMY_NODE_ID,
             kind: ast::StmtKind::Expr(DummyResult::raw_expr(self.span, self.is_error)),
             span: self.span,
-            tokens: None
         }])
     }
 

compiler/rustc_expand/src/build.rs

@@ -140,12 +140,7 @@ impl<'a> ExtCtxt<'a> {
     }
 
     pub fn stmt_expr(&self, expr: P<ast::Expr>) -> ast::Stmt {
-        ast::Stmt {
-            id: ast::DUMMY_NODE_ID,
-            span: expr.span,
-            kind: ast::StmtKind::Expr(expr),
-            tokens: None,
-        }
+        ast::Stmt { id: ast::DUMMY_NODE_ID, span: expr.span, kind: ast::StmtKind::Expr(expr) }
     }
 
     pub fn stmt_let(&self, sp: Span, mutbl: bool, ident: Ident, ex: P<ast::Expr>) -> ast::Stmt {
@@ -162,13 +157,9 @@ impl<'a> ExtCtxt<'a> {
             id: ast::DUMMY_NODE_ID,
             span: sp,
             attrs: AttrVec::new(),
-        });
-        ast::Stmt {
-            id: ast::DUMMY_NODE_ID,
-            kind: ast::StmtKind::Local(local),
-            span: sp,
             tokens: None,
-        }
+        });
+        ast::Stmt { id: ast::DUMMY_NODE_ID, kind: ast::StmtKind::Local(local), span: sp }
     }
 
     // Generates `let _: Type;`, which is usually used for type assertions.
@@ -180,17 +171,13 @@ impl<'a> ExtCtxt<'a> {
             id: ast::DUMMY_NODE_ID,
             span,
             attrs: AttrVec::new(),
+            tokens: None,
         });
-        ast::Stmt { id: ast::DUMMY_NODE_ID, kind: ast::StmtKind::Local(local), span, tokens: None }
+        ast::Stmt { id: ast::DUMMY_NODE_ID, kind: ast::StmtKind::Local(local), span }
     }
 
     pub fn stmt_item(&self, sp: Span, item: P<ast::Item>) -> ast::Stmt {
-        ast::Stmt {
-            id: ast::DUMMY_NODE_ID,
-            kind: ast::StmtKind::Item(item),
-            span: sp,
-            tokens: None,
-        }
+        ast::Stmt { id: ast::DUMMY_NODE_ID, kind: ast::StmtKind::Item(item), span: sp }
     }
 
     pub fn block_expr(&self, expr: P<ast::Expr>) -> P<ast::Block> {
@@ -200,7 +187,6 @@ impl<'a> ExtCtxt<'a> {
                 id: ast::DUMMY_NODE_ID,
                 span: expr.span,
                 kind: ast::StmtKind::Expr(expr),
-                tokens: None,
             }],
         )
     }

compiler/rustc_expand/src/expand.rs

@@ -1274,12 +1274,6 @@ impl<'a, 'b> MutVisitor for InvocationCollector<'a, 'b> {
         // we'll expand attributes on expressions separately
         if !stmt.is_expr() {
             let attr = if stmt.is_item() {
-                // FIXME: Implement proper token collection for statements
-                if let StmtKind::Item(item) = &mut stmt.kind {
-                    stmt.tokens = item.tokens.take()
-                } else {
-                    unreachable!()
-                };
                 self.take_first_attr(&mut stmt)
             } else {
                 // Ignore derives on non-item statements for backwards compatibility.
@@ -1295,7 +1289,7 @@ impl<'a, 'b> MutVisitor for InvocationCollector<'a, 'b> {
         }
 
         if let StmtKind::MacCall(mac) = stmt.kind {
-            let MacCallStmt { mac, style, attrs } = mac.into_inner();
+            let MacCallStmt { mac, style, attrs, tokens: _ } = mac.into_inner();
             self.check_attributes(&attrs);
             let mut placeholder =
                 self.collect_bang(mac, stmt.span, AstFragmentKind::Stmts).make_stmts();
@@ -1312,10 +1306,10 @@ impl<'a, 'b> MutVisitor for InvocationCollector<'a, 'b> {
         }
 
         // The placeholder expander gives ids to statements, so we avoid folding the id here.
-        let ast::Stmt { id, kind, span, tokens } = stmt;
+        let ast::Stmt { id, kind, span } = stmt;
         noop_flat_map_stmt_kind(kind, self)
             .into_iter()
-            .map(|kind| ast::Stmt { id, kind, span, tokens: tokens.clone() })
+            .map(|kind| ast::Stmt { id, kind, span })
             .collect()
     }
 

compiler/rustc_expand/src/placeholders.rs

@@ -104,8 +104,9 @@ pub fn placeholder(
                 mac: mac_placeholder(),
                 style: ast::MacStmtStyle::Braces,
                 attrs: ast::AttrVec::new(),
+                tokens: None,
             });
-            ast::Stmt { id, span, kind: ast::StmtKind::MacCall(mac), tokens: None }
+            ast::Stmt { id, span, kind: ast::StmtKind::MacCall(mac) }
         }]),
         AstFragmentKind::Arms => AstFragment::Arms(smallvec![ast::Arm {
             attrs: Default::default(),
@@ -331,12 +332,8 @@ impl<'a, 'b> MutVisitor for PlaceholderExpander<'a, 'b> {
 
             // FIXME: We will need to preserve the original semicolon token and
             // span as part of #15701
-            let empty_stmt = ast::Stmt {
-                id: ast::DUMMY_NODE_ID,
-                kind: ast::StmtKind::Empty,
-                span: DUMMY_SP,
-                tokens: None,
-            };
+            let empty_stmt =
+                ast::Stmt { id: ast::DUMMY_NODE_ID, kind: ast::StmtKind::Empty, span: DUMMY_SP };
 
             if let Some(stmt) = stmts.pop() {
                 if stmt.has_trailing_semicolon() {

compiler/rustc_interface/src/util.rs

@@ -810,7 +810,6 @@ impl<'a> MutVisitor for ReplaceBodyWithLoop<'a, '_> {
                 id: resolver.next_node_id(),
                 kind: ast::StmtKind::Expr(expr),
                 span: rustc_span::DUMMY_SP,
-                tokens: None,
             }
         }
 
@@ -827,7 +826,6 @@ impl<'a> MutVisitor for ReplaceBodyWithLoop<'a, '_> {
             id: self.resolver.next_node_id(),
             span: rustc_span::DUMMY_SP,
             kind: ast::StmtKind::Expr(loop_expr),
-            tokens: None,
         };
 
         if self.within_static_or_const {

compiler/rustc_parse/src/lib.rs

@@ -6,6 +6,7 @@
 #![feature(or_patterns)]
 
 use rustc_ast as ast;
+use rustc_ast::attr::HasAttrs;
 use rustc_ast::token::{self, DelimToken, Nonterminal, Token, TokenKind};
 use rustc_ast::tokenstream::{self, LazyTokenStream, TokenStream, TokenTree};
 use rustc_ast_pretty::pprust;
@@ -249,29 +250,23 @@ pub fn nt_to_tokenstream(nt: &Nonterminal, sess: &ParseSess, span: Span) -> Toke
     // before we fall back to the stringification.
 
     let convert_tokens =
-        |tokens: &Option<LazyTokenStream>| tokens.as_ref().map(|t| t.create_token_stream());
+        |tokens: Option<&LazyTokenStream>| tokens.as_ref().map(|t| t.create_token_stream());
 
     let tokens = match *nt {
         Nonterminal::NtItem(ref item) => prepend_attrs(&item.attrs, item.tokens.as_ref()),
-        Nonterminal::NtBlock(ref block) => convert_tokens(&block.tokens),
-        Nonterminal::NtStmt(ref stmt) => {
-            // FIXME: We currently only collect tokens for `:stmt`
-            // matchers in `macro_rules!` macros. When we start collecting
-            // tokens for attributes on statements, we will need to prepend
-            // attributes here
-            convert_tokens(&stmt.tokens)
-        }
-        Nonterminal::NtPat(ref pat) => convert_tokens(&pat.tokens),
-        Nonterminal::NtTy(ref ty) => convert_tokens(&ty.tokens),
+        Nonterminal::NtBlock(ref block) => convert_tokens(block.tokens.as_ref()),
+        Nonterminal::NtStmt(ref stmt) => prepend_attrs(stmt.attrs(), stmt.tokens()),
+        Nonterminal::NtPat(ref pat) => convert_tokens(pat.tokens.as_ref()),
+        Nonterminal::NtTy(ref ty) => convert_tokens(ty.tokens.as_ref()),
         Nonterminal::NtIdent(ident, is_raw) => {
             Some(tokenstream::TokenTree::token(token::Ident(ident.name, is_raw), ident.span).into())
         }
         Nonterminal::NtLifetime(ident) => {
             Some(tokenstream::TokenTree::token(token::Lifetime(ident.name), ident.span).into())
         }
-        Nonterminal::NtMeta(ref attr) => convert_tokens(&attr.tokens),
-        Nonterminal::NtPath(ref path) => convert_tokens(&path.tokens),
-        Nonterminal::NtVis(ref vis) => convert_tokens(&vis.tokens),
+        Nonterminal::NtMeta(ref attr) => convert_tokens(attr.tokens.as_ref()),
+        Nonterminal::NtPath(ref path) => convert_tokens(path.tokens.as_ref()),
+        Nonterminal::NtVis(ref vis) => convert_tokens(vis.tokens.as_ref()),
         Nonterminal::NtTT(ref tt) => Some(tt.clone().into()),
         Nonterminal::NtExpr(ref expr) | Nonterminal::NtLiteral(ref expr) => {
             if expr.tokens.is_none() {