diff options
| author | Tim Diekmann <tim.diekmann@3dvision.de> | 2020-10-25 16:32:28 +0100 |
|---|---|---|
| committer | Tim Diekmann <tim.diekmann@3dvision.de> | 2020-10-25 16:32:28 +0100 |
| commit | 06e4497a04615ad95dff4240ca9980f19ed364ad (patch) | |
| tree | 123a17d04628e05bb2448d1d3e3f11e60f240304 /compiler/rustc_parse/src/parser | |
| parent | 693a2bf18b7090202784f561de3dfca45c4f79be (diff) | |
| parent | f392479de6b003e72f93cb8f9955b3cf4135c2cd (diff) | |
| download | rust-06e4497a04615ad95dff4240ca9980f19ed364ad.tar.gz rust-06e4497a04615ad95dff4240ca9980f19ed364ad.zip | |
Merge remote-tracking branch 'upstream/master' into box-alloc
Diffstat (limited to 'compiler/rustc_parse/src/parser')
| -rw-r--r-- | compiler/rustc_parse/src/parser/attr.rs | 150 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/parser/expr.rs | 23 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/parser/item.rs | 7 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/parser/mod.rs | 312 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/parser/nonterminal.rs | 22 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/parser/pat.rs | 20 |
6 files changed, 312 insertions, 222 deletions
diff --git a/compiler/rustc_parse/src/parser/attr.rs b/compiler/rustc_parse/src/parser/attr.rs index 98f94098bfc..053b7e0b75f 100644 --- a/compiler/rustc_parse/src/parser/attr.rs +++ b/compiler/rustc_parse/src/parser/attr.rs @@ -4,7 +4,7 @@ use rustc_ast::attr; use rustc_ast::token::{self, Nonterminal}; use rustc_ast_pretty::pprust; use rustc_errors::{error_code, PResult}; -use rustc_span::Span; +use rustc_span::{sym, Span}; use tracing::debug; @@ -30,41 +30,53 @@ impl<'a> Parser<'a> { let mut just_parsed_doc_comment = false; loop { debug!("parse_outer_attributes: self.token={:?}", self.token); - if self.check(&token::Pound) { - let inner_error_reason = if just_parsed_doc_comment { - "an inner attribute is not permitted following an outer doc comment" - } else if !attrs.is_empty() { - "an inner attribute is not permitted following an outer attribute" - } else { - DEFAULT_UNEXPECTED_INNER_ATTR_ERR_MSG - }; - let inner_parse_policy = InnerAttrPolicy::Forbidden { - reason: inner_error_reason, - saw_doc_comment: just_parsed_doc_comment, - prev_attr_sp: attrs.last().map(|a| a.span), - }; - let attr = self.parse_attribute_with_inner_parse_policy(inner_parse_policy)?; - attrs.push(attr); - just_parsed_doc_comment = false; + let (attr, tokens) = if self.check(&token::Pound) { + self.collect_tokens(|this| { + let inner_error_reason = if just_parsed_doc_comment { + "an inner attribute is not permitted following an outer doc comment" + } else if !attrs.is_empty() { + "an inner attribute is not permitted following an outer attribute" + } else { + DEFAULT_UNEXPECTED_INNER_ATTR_ERR_MSG + }; + let inner_parse_policy = InnerAttrPolicy::Forbidden { + reason: inner_error_reason, + saw_doc_comment: just_parsed_doc_comment, + prev_attr_sp: attrs.last().map(|a| a.span), + }; + let attr = this.parse_attribute_with_inner_parse_policy(inner_parse_policy)?; + just_parsed_doc_comment = false; + Ok(Some(attr)) + })? } else if let token::DocComment(comment_kind, attr_style, data) = self.token.kind { - let attr = attr::mk_doc_comment(comment_kind, attr_style, data, self.token.span); - if attr.style != ast::AttrStyle::Outer { - self.sess - .span_diagnostic - .struct_span_err_with_code( - self.token.span, - "expected outer doc comment", - error_code!(E0753), - ) - .note( - "inner doc comments like this (starting with \ - `//!` or `/*!`) can only appear before items", - ) - .emit(); - } + self.collect_tokens(|this| { + let attr = + attr::mk_doc_comment(comment_kind, attr_style, data, this.token.span); + if attr.style != ast::AttrStyle::Outer { + this.sess + .span_diagnostic + .struct_span_err_with_code( + this.token.span, + "expected outer doc comment", + error_code!(E0753), + ) + .note( + "inner doc comments like this (starting with \ + `//!` or `/*!`) can only appear before items", + ) + .emit(); + } + this.bump(); + just_parsed_doc_comment = true; + Ok(Some(attr)) + })? + } else { + (None, None) + }; + + if let Some(mut attr) = attr { + attr.tokens = tokens; attrs.push(attr); - self.bump(); - just_parsed_doc_comment = true; } else { break; } @@ -99,7 +111,7 @@ impl<'a> Parser<'a> { if self.eat(&token::Not) { ast::AttrStyle::Inner } else { ast::AttrStyle::Outer }; self.expect(&token::OpenDelim(token::Bracket))?; - let item = self.parse_attr_item()?; + let item = self.parse_attr_item(false)?; self.expect(&token::CloseDelim(token::Bracket))?; let attr_sp = lo.to(self.prev_token.span); @@ -148,7 +160,7 @@ impl<'a> Parser<'a> { /// PATH /// PATH `=` UNSUFFIXED_LIT /// The delimiters or `=` are still put into the resulting token stream. - pub fn parse_attr_item(&mut self) -> PResult<'a, ast::AttrItem> { + pub fn parse_attr_item(&mut self, capture_tokens: bool) -> PResult<'a, ast::AttrItem> { let item = match self.token.kind { token::Interpolated(ref nt) => match **nt { Nonterminal::NtMeta(ref item) => Some(item.clone().into_inner()), @@ -160,9 +172,18 @@ impl<'a> Parser<'a> { self.bump(); item } else { - let path = self.parse_path(PathStyle::Mod)?; - let args = self.parse_attr_args()?; - ast::AttrItem { path, args, tokens: None } + let do_parse = |this: &mut Self| { + let path = this.parse_path(PathStyle::Mod)?; + let args = this.parse_attr_args()?; + Ok(ast::AttrItem { path, args, tokens: None }) + }; + if capture_tokens { + let (mut item, tokens) = self.collect_tokens(do_parse)?; + item.tokens = tokens; + item + } else { + do_parse(self)? + } }) } @@ -175,19 +196,31 @@ impl<'a> Parser<'a> { let mut attrs: Vec<ast::Attribute> = vec![]; loop { // Only try to parse if it is an inner attribute (has `!`). - if self.check(&token::Pound) && self.look_ahead(1, |t| t == &token::Not) { - let attr = self.parse_attribute(true)?; - assert_eq!(attr.style, ast::AttrStyle::Inner); - attrs.push(attr); - } else if let token::DocComment(comment_kind, attr_style, data) = self.token.kind { - // We need to get the position of this token before we bump. - let attr = attr::mk_doc_comment(comment_kind, attr_style, data, self.token.span); - if attr.style == ast::AttrStyle::Inner { - attrs.push(attr); - self.bump(); + let (attr, tokens) = + if self.check(&token::Pound) && self.look_ahead(1, |t| t == &token::Not) { + self.collect_tokens(|this| { + let attr = this.parse_attribute(true)?; + assert_eq!(attr.style, ast::AttrStyle::Inner); + Ok(Some(attr)) + })? + } else if let token::DocComment(comment_kind, attr_style, data) = self.token.kind { + self.collect_tokens(|this| { + // We need to get the position of this token before we bump. + let attr = + attr::mk_doc_comment(comment_kind, attr_style, data, this.token.span); + if attr.style == ast::AttrStyle::Inner { + this.bump(); + Ok(Some(attr)) + } else { + Ok(None) + } + })? } else { - break; - } + (None, None) + }; + if let Some(mut attr) = attr { + attr.tokens = tokens; + attrs.push(attr); } else { break; } @@ -220,7 +253,7 @@ impl<'a> Parser<'a> { let mut expanded_attrs = Vec::with_capacity(1); while self.token.kind != token::Eof { let lo = self.token.span; - let item = self.parse_attr_item()?; + let item = self.parse_attr_item(true)?; expanded_attrs.push((item, lo.to(self.prev_token.span))); if !self.eat(&token::Comma) { break; @@ -302,3 +335,16 @@ impl<'a> Parser<'a> { Err(self.struct_span_err(self.token.span, &msg)) } } + +pub fn maybe_needs_tokens(attrs: &[ast::Attribute]) -> bool { + attrs.iter().any(|attr| { + if let Some(ident) = attr.ident() { + ident.name == sym::derive + // This might apply a custom attribute/derive + || ident.name == sym::cfg_attr + || !rustc_feature::is_builtin_attr_name(ident.name) + } else { + true + } + }) +} diff --git a/compiler/rustc_parse/src/parser/expr.rs b/compiler/rustc_parse/src/parser/expr.rs index 17cbaf65420..c44e00f861d 100644 --- a/compiler/rustc_parse/src/parser/expr.rs +++ b/compiler/rustc_parse/src/parser/expr.rs @@ -6,6 +6,7 @@ use crate::maybe_recover_from_interpolated_ty_qpath; use rustc_ast::ptr::P; use rustc_ast::token::{self, Token, TokenKind}; +use rustc_ast::tokenstream::Spacing; use rustc_ast::util::classify; use rustc_ast::util::literal::LitError; use rustc_ast::util::parser::{prec_let_scrutinee_needs_par, AssocOp, Fixity}; @@ -18,7 +19,6 @@ use rustc_span::source_map::{self, Span, Spanned}; use rustc_span::symbol::{kw, sym, Ident, Symbol}; use rustc_span::{BytePos, Pos}; use std::mem; -use tracing::debug; /// Possibly accepts an `token::Interpolated` expression (a pre-parsed expression /// dropped into the token stream, which happens while parsing the result of @@ -459,7 +459,7 @@ impl<'a> Parser<'a> { /// Parses a prefix-unary-operator expr. fn parse_prefix_expr(&mut self, attrs: Option<AttrVec>) -> PResult<'a, P<Expr>> { let attrs = self.parse_or_use_outer_attributes(attrs)?; - self.maybe_collect_tokens(!attrs.is_empty(), |this| { + self.maybe_collect_tokens(super::attr::maybe_needs_tokens(&attrs), |this| { let lo = this.token.span; // Note: when adding new unary operators, don't forget to adjust TokenKind::can_begin_expr() let (hi, ex) = match this.token.uninterpolate().kind { @@ -884,7 +884,7 @@ impl<'a> Parser<'a> { assert!(suffix.is_none()); let symbol = Symbol::intern(&i); self.token = Token::new(token::Ident(symbol, false), ident_span); - let next_token = Token::new(token::Dot, dot_span); + let next_token = (Token::new(token::Dot, dot_span), self.token_spacing); self.parse_tuple_field_access_expr(lo, base, symbol, None, Some(next_token)) } // 1.2 | 1.2e3 @@ -902,12 +902,14 @@ impl<'a> Parser<'a> { }; let symbol1 = Symbol::intern(&i1); self.token = Token::new(token::Ident(symbol1, false), ident1_span); - let next_token1 = Token::new(token::Dot, dot_span); + // This needs to be `Spacing::Alone` to prevent regressions. + // See issue #76399 and PR #76285 for more details + let next_token1 = (Token::new(token::Dot, dot_span), Spacing::Alone); let base1 = self.parse_tuple_field_access_expr(lo, base, symbol1, None, Some(next_token1)); let symbol2 = Symbol::intern(&i2); let next_token2 = Token::new(token::Ident(symbol2, false), ident2_span); - self.bump_with(next_token2); // `.` + self.bump_with((next_token2, self.token_spacing)); // `.` self.parse_tuple_field_access_expr(lo, base1, symbol2, suffix, None) } // 1e+ | 1e- (recovered) @@ -930,7 +932,7 @@ impl<'a> Parser<'a> { base: P<Expr>, field: Symbol, suffix: Option<Symbol>, - next_token: Option<Token>, + next_token: Option<(Token, Spacing)>, ) -> P<Expr> { match next_token { Some(next_token) => self.bump_with(next_token), @@ -1060,6 +1062,8 @@ impl<'a> Parser<'a> { }) } else if self.eat_keyword(kw::Unsafe) { self.parse_block_expr(None, lo, BlockCheckMode::Unsafe(ast::UserProvided), attrs) + } else if self.check_inline_const(0) { + self.parse_const_block(lo.to(self.token.span)) } else if self.is_do_catch_block() { self.recover_do_catch(attrs) } else if self.is_try_block() { @@ -1107,13 +1111,12 @@ impl<'a> Parser<'a> { fn maybe_collect_tokens( &mut self, - has_outer_attrs: bool, + needs_tokens: bool, f: impl FnOnce(&mut Self) -> PResult<'a, P<Expr>>, ) -> PResult<'a, P<Expr>> { - if has_outer_attrs { + if needs_tokens { let (mut expr, tokens) = self.collect_tokens(f)?; - debug!("maybe_collect_tokens: Collected tokens for {:?} (tokens {:?}", expr, tokens); - expr.tokens = Some(tokens); + expr.tokens = tokens; Ok(expr) } else { f(self) diff --git a/compiler/rustc_parse/src/parser/item.rs b/compiler/rustc_parse/src/parser/item.rs index 48341f71d33..e57a2e42b5d 100644 --- a/compiler/rustc_parse/src/parser/item.rs +++ b/compiler/rustc_parse/src/parser/item.rs @@ -116,15 +116,16 @@ impl<'a> Parser<'a> { Some(item.into_inner()) }); + let needs_tokens = super::attr::maybe_needs_tokens(&attrs); + let mut unclosed_delims = vec![]; - let has_attrs = !attrs.is_empty(); let parse_item = |this: &mut Self| { let item = this.parse_item_common_(attrs, mac_allowed, attrs_allowed, req_name); unclosed_delims.append(&mut this.unclosed_delims); item }; - let (mut item, tokens) = if has_attrs { + let (mut item, tokens) = if needs_tokens { let (item, tokens) = self.collect_tokens(parse_item)?; (item, Some(tokens)) } else { @@ -150,7 +151,7 @@ impl<'a> Parser<'a> { if let Some(tokens) = tokens { if let Some(item) = &mut item { if !item.attrs.iter().any(|attr| attr.style == AttrStyle::Inner) { - item.tokens = Some(tokens); + item.tokens = tokens; } } } diff --git a/compiler/rustc_parse/src/parser/mod.rs b/compiler/rustc_parse/src/parser/mod.rs index c1094681221..175dd3fa53a 100644 --- a/compiler/rustc_parse/src/parser/mod.rs +++ b/compiler/rustc_parse/src/parser/mod.rs @@ -16,12 +16,15 @@ pub use path::PathStyle; use rustc_ast::ptr::P; use rustc_ast::token::{self, DelimToken, Token, TokenKind}; -use rustc_ast::tokenstream::{self, DelimSpan, TokenStream, TokenTree, TreeAndSpacing}; +use rustc_ast::tokenstream::{self, DelimSpan, LazyTokenStream, LazyTokenStreamInner, Spacing}; +use rustc_ast::tokenstream::{TokenStream, TokenTree}; use rustc_ast::DUMMY_NODE_ID; -use rustc_ast::{self as ast, AttrStyle, AttrVec, Const, CrateSugar, Extern, Unsafe}; -use rustc_ast::{Async, MacArgs, MacDelimiter, Mutability, StrLit, Visibility, VisibilityKind}; +use rustc_ast::{self as ast, AnonConst, AttrStyle, AttrVec, Const, CrateSugar, Extern, Unsafe}; +use rustc_ast::{Async, Expr, ExprKind, MacArgs, MacDelimiter, Mutability, StrLit}; +use rustc_ast::{Visibility, VisibilityKind}; use rustc_ast_pretty::pprust; -use rustc_errors::{struct_span_err, Applicability, DiagnosticBuilder, FatalError, PResult}; +use rustc_errors::PResult; +use rustc_errors::{struct_span_err, Applicability, DiagnosticBuilder, FatalError}; use rustc_session::parse::ParseSess; use rustc_span::source_map::{Span, DUMMY_SP}; use rustc_span::symbol::{kw, sym, Ident, Symbol}; @@ -84,10 +87,14 @@ pub struct Parser<'a> { pub sess: &'a ParseSess, /// The current token. pub token: Token, + /// The spacing for the current token + pub token_spacing: Spacing, /// The previous token. pub prev_token: Token, restrictions: Restrictions, expected_tokens: Vec<TokenType>, + // Important: This must only be advanced from `next_tok` + // to ensure that `token_cursor.num_next_calls` is updated properly token_cursor: TokenCursor, desugar_doc_comments: bool, /// This field is used to keep track of how many left angle brackets we have seen. This is @@ -119,8 +126,10 @@ impl<'a> Drop for Parser<'a> { struct TokenCursor { frame: TokenCursorFrame, stack: Vec<TokenCursorFrame>, - cur_token: Option<TreeAndSpacing>, - collecting: Option<Collecting>, + desugar_doc_comments: bool, + // Counts the number of calls to `next` or `next_desugared`, + // depending on whether `desugar_doc_comments` is set. + num_next_calls: usize, } #[derive(Clone)] @@ -132,40 +141,22 @@ struct TokenCursorFrame { close_delim: bool, } -/// Used to track additional state needed by `collect_tokens` -#[derive(Clone, Debug)] -struct Collecting { - /// Holds the current tokens captured during the most - /// recent call to `collect_tokens` - buf: Vec<TreeAndSpacing>, - /// The depth of the `TokenCursor` stack at the time - /// collection was started. When we encounter a `TokenTree::Delimited`, - /// we want to record the `TokenTree::Delimited` itself, - /// but *not* any of the inner tokens while we are inside - /// the new frame (this would cause us to record duplicate tokens). - /// - /// This `depth` fields tracks stack depth we are recording tokens. - /// Only tokens encountered at this depth will be recorded. See - /// `TokenCursor::next` for more details. - depth: usize, -} - impl TokenCursorFrame { - fn new(span: DelimSpan, delim: DelimToken, tts: &TokenStream) -> Self { + fn new(span: DelimSpan, delim: DelimToken, tts: TokenStream) -> Self { TokenCursorFrame { delim, span, open_delim: delim == token::NoDelim, - tree_cursor: tts.clone().into_trees(), + tree_cursor: tts.into_trees(), close_delim: delim == token::NoDelim, } } } impl TokenCursor { - fn next(&mut self) -> Token { + fn next(&mut self) -> (Token, Spacing) { loop { - let tree = if !self.frame.open_delim { + let (tree, spacing) = if !self.frame.open_delim { self.frame.open_delim = true; TokenTree::open_tt(self.frame.span, self.frame.delim).into() } else if let Some(tree) = self.frame.tree_cursor.next_with_spacing() { @@ -177,40 +168,24 @@ impl TokenCursor { self.frame = frame; continue; } else { - return Token::new(token::Eof, DUMMY_SP); + (TokenTree::Token(Token::new(token::Eof, DUMMY_SP)), Spacing::Alone) }; - // Don't set an open delimiter as our current token - we want - // to leave it as the full `TokenTree::Delimited` from the previous - // iteration of this loop - if !matches!(tree.0, TokenTree::Token(Token { kind: TokenKind::OpenDelim(_), .. })) { - self.cur_token = Some(tree.clone()); - } - - if let Some(collecting) = &mut self.collecting { - if collecting.depth == self.stack.len() { - debug!( - "TokenCursor::next(): collected {:?} at depth {:?}", - tree, - self.stack.len() - ); - collecting.buf.push(tree.clone()) + match tree { + TokenTree::Token(token) => { + return (token, spacing); } - } - - match tree.0 { - TokenTree::Token(token) => return token, TokenTree::Delimited(sp, delim, tts) => { - let frame = TokenCursorFrame::new(sp, delim, &tts); + let frame = TokenCursorFrame::new(sp, delim, tts); self.stack.push(mem::replace(&mut self.frame, frame)); } } } } - fn next_desugared(&mut self) -> Token { + fn next_desugared(&mut self) -> (Token, Spacing) { let (data, attr_style, sp) = match self.next() { - Token { kind: token::DocComment(_, attr_style, data), span } => { + (Token { kind: token::DocComment(_, attr_style, data), span }, _) => { (data, attr_style, span) } tok => return tok, @@ -248,7 +223,7 @@ impl TokenCursor { TokenCursorFrame::new( delim_span, token::NoDelim, - &if attr_style == AttrStyle::Inner { + if attr_style == AttrStyle::Inner { [TokenTree::token(token::Pound, sp), TokenTree::token(token::Not, sp), body] .iter() .cloned() @@ -350,14 +325,15 @@ impl<'a> Parser<'a> { let mut parser = Parser { sess, token: Token::dummy(), + token_spacing: Spacing::Alone, prev_token: Token::dummy(), restrictions: Restrictions::empty(), expected_tokens: Vec::new(), token_cursor: TokenCursor { - frame: TokenCursorFrame::new(DelimSpan::dummy(), token::NoDelim, &tokens), + frame: TokenCursorFrame::new(DelimSpan::dummy(), token::NoDelim, tokens), stack: Vec::new(), - cur_token: None, - collecting: None, + num_next_calls: 0, + desugar_doc_comments, }, desugar_doc_comments, unmatched_angle_bracket_count: 0, @@ -374,17 +350,18 @@ impl<'a> Parser<'a> { parser } - fn next_tok(&mut self, fallback_span: Span) -> Token { - let mut next = if self.desugar_doc_comments { + fn next_tok(&mut self, fallback_span: Span) -> (Token, Spacing) { + let (mut next, spacing) = if self.desugar_doc_comments { self.token_cursor.next_desugared() } else { self.token_cursor.next() }; + self.token_cursor.num_next_calls += 1; if next.span.is_dummy() { // Tweak the location for better diagnostics, but keep syntactic context intact. next.span = fallback_span.with_ctxt(next.span.ctxt()); } - next + (next, spacing) } pub fn unexpected<T>(&mut self) -> PResult<'a, T> { @@ -545,6 +522,15 @@ impl<'a> Parser<'a> { self.check_or_expected(self.token.can_begin_const_arg(), TokenType::Const) } + fn check_inline_const(&self, dist: usize) -> bool { + self.is_keyword_ahead(dist, &[kw::Const]) + && self.look_ahead(dist + 1, |t| match t.kind { + token::Interpolated(ref nt) => matches!(**nt, token::NtBlock(..)), + token::OpenDelim(DelimToken::Brace) => true, + _ => false, + }) + } + /// Checks to see if the next token is either `+` or `+=`. /// Otherwise returns `false`. fn check_plus(&mut self) -> bool { @@ -567,7 +553,9 @@ impl<'a> Parser<'a> { let first_span = self.sess.source_map().start_point(self.token.span); let second_span = self.token.span.with_lo(first_span.hi()); self.token = Token::new(first, first_span); - self.bump_with(Token::new(second, second_span)); + // Use the spacing of the glued token as the spacing + // of the unglued second token. + self.bump_with((Token::new(second, second_span), self.token_spacing)); true } _ => { @@ -799,7 +787,7 @@ impl<'a> Parser<'a> { } /// Advance the parser by one token using provided token as the next one. - fn bump_with(&mut self, next_token: Token) { + fn bump_with(&mut self, (next_token, next_spacing): (Token, Spacing)) { // Bumping after EOF is a bad sign, usually an infinite loop. if self.prev_token.kind == TokenKind::Eof { let msg = "attempted to bump the parser past EOF (may be stuck in a loop)"; @@ -808,6 +796,7 @@ impl<'a> Parser<'a> { // Update the current and previous tokens. self.prev_token = mem::replace(&mut self.token, next_token); + self.token_spacing = next_spacing; // Diagnostics. self.expected_tokens.clear(); @@ -827,15 +816,15 @@ impl<'a> Parser<'a> { } let frame = &self.token_cursor.frame; - looker(&match frame.tree_cursor.look_ahead(dist - 1) { + match frame.tree_cursor.look_ahead(dist - 1) { Some(tree) => match tree { - TokenTree::Token(token) => token, + TokenTree::Token(token) => looker(token), TokenTree::Delimited(dspan, delim, _) => { - Token::new(token::OpenDelim(delim), dspan.open) + looker(&Token::new(token::OpenDelim(*delim), dspan.open)) } }, - None => Token::new(token::CloseDelim(frame.delim), frame.span.close), - }) + None => looker(&Token::new(token::CloseDelim(frame.delim), frame.span.close)), + } } /// Returns whether any of the given keywords are `dist` tokens ahead of the current one. @@ -864,13 +853,28 @@ impl<'a> Parser<'a> { /// Parses constness: `const` or nothing. fn parse_constness(&mut self) -> Const { - if self.eat_keyword(kw::Const) { + // Avoid const blocks to be parsed as const items + if self.look_ahead(1, |t| t != &token::OpenDelim(DelimToken::Brace)) + && self.eat_keyword(kw::Const) + { Const::Yes(self.prev_token.uninterpolated_span()) } else { Const::No } } + /// Parses inline const expressions. + fn parse_const_block(&mut self, span: Span) -> PResult<'a, P<Expr>> { + self.sess.gated_spans.gate(sym::inline_const, span); + self.eat_keyword(kw::Const); + let blk = self.parse_block()?; + let anon_const = AnonConst { + id: DUMMY_NODE_ID, + value: self.mk_expr(blk.span, ExprKind::Block(blk, None), AttrVec::new()), + }; + Ok(self.mk_expr(span, ExprKind::ConstBlock(anon_const), AttrVec::new())) + } + /// Parses mutability (`mut` or nothing). fn parse_mutability(&mut self) -> Mutability { if self.eat_keyword(kw::Mut) { Mutability::Mut } else { Mutability::Not } @@ -963,13 +967,27 @@ impl<'a> Parser<'a> { pub(crate) fn parse_token_tree(&mut self) -> TokenTree { match self.token.kind { token::OpenDelim(..) => { - let frame = mem::replace( - &mut self.token_cursor.frame, - self.token_cursor.stack.pop().unwrap(), - ); - self.token = Token::new(TokenKind::CloseDelim(frame.delim), frame.span.close); + let depth = self.token_cursor.stack.len(); + + // We keep advancing the token cursor until we hit + // the matching `CloseDelim` token. + while !(depth == self.token_cursor.stack.len() + && matches!(self.token.kind, token::CloseDelim(_))) + { + // Advance one token at a time, so `TokenCursor::next()` + // can capture these tokens if necessary. + self.bump(); + } + // We are still inside the frame corresponding + // to the delimited stream we captured, so grab + // the tokens from this frame. + let frame = &self.token_cursor.frame; + let stream = frame.tree_cursor.stream.clone(); + let span = frame.span; + let delim = frame.delim; + // Consume close delimiter self.bump(); - TokenTree::Delimited(frame.span, frame.delim, frame.tree_cursor.stream) + TokenTree::Delimited(span, delim, stream) } token::CloseDelim(_) | token::Eof => unreachable!(), _ => { @@ -1160,8 +1178,9 @@ impl<'a> Parser<'a> { /// Records all tokens consumed by the provided callback, /// including the current token. These tokens are collected - /// into a `TokenStream`, and returned along with the result - /// of the callback. + /// into a `LazyTokenStream`, and returned along with the result + /// of the callback. The returned `LazyTokenStream` will be `None` + /// if not tokens were captured. /// /// Note: If your callback consumes an opening delimiter /// (including the case where you call `collect_tokens` @@ -1177,79 +1196,50 @@ impl<'a> Parser<'a> { pub fn collect_tokens<R>( &mut self, f: impl FnOnce(&mut Self) -> PResult<'a, R>, - ) -> PResult<'a, (R, TokenStream)> { - // Record all tokens we parse when parsing this item. - let tokens: Vec<TreeAndSpacing> = self.token_cursor.cur_token.clone().into_iter().collect(); - debug!("collect_tokens: starting with {:?}", tokens); - - // We need special handling for the case where `collect_tokens` is called - // on an opening delimeter (e.g. '('). At this point, we have already pushed - // a new frame - however, we want to record the original `TokenTree::Delimited`, - // for consistency with the case where we start recording one token earlier. - // See `TokenCursor::next` to see how `cur_token` is set up. - let prev_depth = - if matches!(self.token_cursor.cur_token, Some((TokenTree::Delimited(..), _))) { - if self.token_cursor.stack.is_empty() { - // There is nothing below us in the stack that - // the function could consume, so the only thing it can legally - // capture is the entire contents of the current frame. - return Ok((f(self)?, TokenStream::new(tokens))); - } - // We have already recorded the full `TokenTree::Delimited` when we created - // our `tokens` vector at the start of this function. We are now inside - // a new frame corresponding to the `TokenTree::Delimited` we already recoreded. - // We don't want to record any of the tokens inside this frame, since they - // will be duplicates of the tokens nested inside the `TokenTree::Delimited`. - // Therefore, we set our recording depth to the *previous* frame. This allows - // us to record a sequence like: `(foo).bar()`: the `(foo)` will be recored - // as our initial `cur_token`, while the `.bar()` will be recored after we - // pop the `(foo)` frame. - self.token_cursor.stack.len() - 1 - } else { - self.token_cursor.stack.len() - }; - let prev_collecting = - self.token_cursor.collecting.replace(Collecting { buf: tokens, depth: prev_depth }); + ) -> PResult<'a, (R, Option<LazyTokenStream>)> { + let start_token = (self.token.clone(), self.token_spacing); + let mut cursor_snapshot = self.token_cursor.clone(); - let ret = f(self); + let ret = f(self)?; - let mut collected_tokens = if let Some(collecting) = self.token_cursor.collecting.take() { - collecting.buf - } else { - let msg = "our vector went away?"; - debug!("collect_tokens: {}", msg); - self.sess.span_diagnostic.delay_span_bug(self.token.span, &msg); - // This can happen due to a bad interaction of two unrelated recovery mechanisms - // with mismatched delimiters *and* recovery lookahead on the likely typo - // `pub ident(` (#62895, different but similar to the case above). - return Ok((ret?, TokenStream::default())); - }; + let new_calls = self.token_cursor.num_next_calls; + let num_calls = new_calls - cursor_snapshot.num_next_calls; + let desugar_doc_comments = self.desugar_doc_comments; - debug!("collect_tokens: got raw tokens {:?}", collected_tokens); - - // If we're not at EOF our current token wasn't actually consumed by - // `f`, but it'll still be in our list that we pulled out. In that case - // put it back. - let extra_token = if self.token != token::Eof { collected_tokens.pop() } else { None }; - - if let Some(mut collecting) = prev_collecting { - // If we were previously collecting at the same depth, - // then the previous call to `collect_tokens` needs to see - // the tokens we just recorded. - // - // If we were previously recording at an lower `depth`, - // then the previous `collect_tokens` call already recorded - // this entire frame in the form of a `TokenTree::Delimited`, - // so there is nothing else for us to do. - if collecting.depth == prev_depth { - collecting.buf.extend(collected_tokens.iter().cloned()); - collecting.buf.extend(extra_token); - debug!("collect_tokens: updating previous buf to {:?}", collecting); - } - self.token_cursor.collecting = Some(collecting) + // We didn't capture any tokens + if num_calls == 0 { + return Ok((ret, None)); } - Ok((ret?, TokenStream::new(collected_tokens))) + // Produces a `TokenStream` on-demand. Using `cursor_snapshot` + // and `num_calls`, we can reconstruct the `TokenStream` seen + // by the callback. This allows us to avoid producing a `TokenStream` + // if it is never needed - for example, a captured `macro_rules!` + // argument that is never passed to a proc macro. + // + // This also makes `Parser` very cheap to clone, since + // there is no intermediate collection buffer to clone. + let lazy_cb = move || { + // The token produced by the final call to `next` or `next_desugared` + // was not actually consumed by the callback. The combination + // of chaining the initial token and using `take` produces the desired + // result - we produce an empty `TokenStream` if no calls were made, + // and omit the final token otherwise. + let tokens = std::iter::once(start_token) + .chain((0..num_calls).map(|_| { + if desugar_doc_comments { + cursor_snapshot.next_desugared() + } else { + cursor_snapshot.next() + } + })) + .take(num_calls); + + make_token_stream(tokens) + }; + let stream = LazyTokenStream::new(LazyTokenStreamInner::Lazy(Box::new(lazy_cb))); + + Ok((ret, Some(stream))) } /// `::{` or `::*` @@ -1298,3 +1288,41 @@ pub fn emit_unclosed_delims(unclosed_delims: &mut Vec<UnmatchedBrace>, sess: &Pa } } } + +/// Converts a flattened iterator of tokens (including open and close delimiter tokens) +/// into a `TokenStream`, creating a `TokenTree::Delimited` for each matching pair +/// of open and close delims. +fn make_token_stream(tokens: impl Iterator<Item = (Token, Spacing)>) -> TokenStream { + #[derive(Debug)] + struct FrameData { + open: Span, + inner: Vec<(TokenTree, Spacing)>, + } + let mut stack = vec![FrameData { open: DUMMY_SP, inner: vec![] }]; + for (token, spacing) in tokens { + match token { + Token { kind: TokenKind::OpenDelim(_), span } => { + stack.push(FrameData { open: span, inner: vec![] }); + } + Token { kind: TokenKind::CloseDelim(delim), span } => { + let frame_data = stack.pop().expect("Token stack was empty!"); + let dspan = DelimSpan::from_pair(frame_data.open, span); + let stream = TokenStream::new(frame_data.inner); + let delimited = TokenTree::Delimited(dspan, delim, stream); + stack + .last_mut() + .unwrap_or_else(|| panic!("Bottom token frame is missing for tokens!")) + .inner + .push((delimited, Spacing::Alone)); + } + token => stack + .last_mut() + .expect("Bottom token frame is missing!") + .inner + .push((TokenTree::Token(token), spacing)), + } + } + let final_buf = stack.pop().expect("Missing final buf!"); + assert!(stack.is_empty(), "Stack should be empty: final_buf={:?} stack={:?}", final_buf, stack); + TokenStream::new(final_buf.inner) +} diff --git a/compiler/rustc_parse/src/parser/nonterminal.rs b/compiler/rustc_parse/src/parser/nonterminal.rs index 15660fd574c..98fb1c82925 100644 --- a/compiler/rustc_parse/src/parser/nonterminal.rs +++ b/compiler/rustc_parse/src/parser/nonterminal.rs @@ -103,7 +103,7 @@ impl<'a> Parser<'a> { // If we captured tokens during parsing (due to outer attributes), // use those. if item.tokens.is_none() { - item.tokens = Some(tokens); + item.tokens = tokens; } token::NtItem(item) } @@ -115,7 +115,7 @@ impl<'a> Parser<'a> { let (mut block, tokens) = self.collect_tokens(|this| this.parse_block())?; // We have have eaten an NtBlock, which could already have tokens if block.tokens.is_none() { - block.tokens = Some(tokens); + block.tokens = tokens; } token::NtBlock(block) } @@ -124,7 +124,7 @@ impl<'a> Parser<'a> { match stmt { Some(mut s) => { if s.tokens.is_none() { - s.tokens = Some(tokens); + s.tokens = tokens; } token::NtStmt(s) } @@ -137,7 +137,7 @@ impl<'a> Parser<'a> { let (mut pat, tokens) = self.collect_tokens(|this| this.parse_pat(None))?; // We have have eaten an NtPat, which could already have tokens if pat.tokens.is_none() { - pat.tokens = Some(tokens); + pat.tokens = tokens; } token::NtPat(pat) } @@ -146,7 +146,7 @@ impl<'a> Parser<'a> { // If we captured tokens during parsing (due to outer attributes), // use those. if expr.tokens.is_none() { - expr.tokens = Some(tokens); + expr.tokens = tokens; } token::NtExpr(expr) } @@ -155,7 +155,7 @@ impl<'a> Parser<'a> { self.collect_tokens(|this| this.parse_literal_maybe_minus())?; // We have have eaten a nonterminal, which could already have tokens if lit.tokens.is_none() { - lit.tokens = Some(tokens); + lit.tokens = tokens; } token::NtLiteral(lit) } @@ -163,7 +163,7 @@ impl<'a> Parser<'a> { let (mut ty, tokens) = self.collect_tokens(|this| this.parse_ty())?; // We have an eaten an NtTy, which could already have tokens if ty.tokens.is_none() { - ty.tokens = Some(tokens); + ty.tokens = tokens; } token::NtTy(ty) } @@ -183,15 +183,15 @@ impl<'a> Parser<'a> { self.collect_tokens(|this| this.parse_path(PathStyle::Type))?; // We have have eaten an NtPath, which could already have tokens if path.tokens.is_none() { - path.tokens = Some(tokens); + path.tokens = tokens; } token::NtPath(path) } NonterminalKind::Meta => { - let (mut attr, tokens) = self.collect_tokens(|this| this.parse_attr_item())?; + let (mut attr, tokens) = self.collect_tokens(|this| this.parse_attr_item(false))?; // We may have eaten a nonterminal, which could already have tokens if attr.tokens.is_none() { - attr.tokens = Some(tokens); + attr.tokens = tokens; } token::NtMeta(P(attr)) } @@ -201,7 +201,7 @@ impl<'a> Parser<'a> { self.collect_tokens(|this| this.parse_visibility(FollowedByType::Yes))?; // We may have etan an `NtVis`, which could already have tokens if vis.tokens.is_none() { - vis.tokens = Some(tokens); + vis.tokens = tokens; } token::NtVis(vis) } diff --git a/compiler/rustc_parse/src/parser/pat.rs b/compiler/rustc_parse/src/parser/pat.rs index 5aced9dc37c..27fe75a23b6 100644 --- a/compiler/rustc_parse/src/parser/pat.rs +++ b/compiler/rustc_parse/src/parser/pat.rs @@ -313,6 +313,15 @@ impl<'a> Parser<'a> { let pat = self.parse_pat_with_range_pat(false, None)?; self.sess.gated_spans.gate(sym::box_patterns, lo.to(self.prev_token.span)); PatKind::Box(pat) + } else if self.check_inline_const(0) { + // Parse `const pat` + let const_expr = self.parse_const_block(lo.to(self.token.span))?; + + if let Some(re) = self.parse_range_end() { + self.parse_pat_range_begin_with(const_expr, re)? + } else { + PatKind::Lit(const_expr) + } } else if self.can_be_ident_pat() { // Parse `ident @ pat` // This can give false positives and parse nullary enums, @@ -714,16 +723,19 @@ impl<'a> Parser<'a> { /// Is the token `dist` away from the current suitable as the start of a range patterns end? fn is_pat_range_end_start(&self, dist: usize) -> bool { - self.look_ahead(dist, |t| { - t.is_path_start() // e.g. `MY_CONST`; + self.check_inline_const(dist) + || self.look_ahead(dist, |t| { + t.is_path_start() // e.g. `MY_CONST`; || t.kind == token::Dot // e.g. `.5` for recovery; || t.can_begin_literal_maybe_minus() // e.g. `42`. || t.is_whole_expr() - }) + }) } fn parse_pat_range_end(&mut self) -> PResult<'a, P<Expr>> { - if self.check_path() { + if self.check_inline_const(0) { + self.parse_const_block(self.token.span) + } else if self.check_path() { let lo = self.token.span; let (qself, path) = if self.eat_lt() { // Parse a qualified path |
