diff --git a/src/librustc_ast/token.rs b/src/librustc_ast/token.rs index 173ea5e48d68..1d26a60f90f4 100644 --- a/src/librustc_ast/token.rs +++ b/src/librustc_ast/token.rs @@ -760,6 +760,44 @@ pub enum Nonterminal { #[cfg(target_arch = "x86_64")] rustc_data_structures::static_assert_size!(Nonterminal, 40); +#[derive(Copy, Clone)] +pub enum NonterminalKind { + Item, + Block, + Stmt, + Pat, + Expr, + Ty, + Ident, + Lifetime, + Literal, + Meta, + Path, + Vis, + TT, +} + +impl NonterminalKind { + pub fn from_symbol(symbol: Symbol) -> Option { + Some(match symbol { + sym::item => NonterminalKind::Item, + sym::block => NonterminalKind::Block, + sym::stmt => NonterminalKind::Stmt, + sym::pat => NonterminalKind::Pat, + sym::expr => NonterminalKind::Expr, + sym::ty => NonterminalKind::Ty, + sym::ident => NonterminalKind::Ident, + sym::lifetime => NonterminalKind::Lifetime, + sym::literal => NonterminalKind::Literal, + sym::meta => NonterminalKind::Meta, + sym::path => NonterminalKind::Path, + sym::vis => NonterminalKind::Vis, + sym::tt => NonterminalKind::TT, + _ => return None, + }) + } +} + impl Nonterminal { fn span(&self) -> Span { match self { diff --git a/src/librustc_expand/mbe/macro_parser.rs b/src/librustc_expand/mbe/macro_parser.rs index 3c15a81c67f6..d0d2e5a2124a 100644 --- a/src/librustc_expand/mbe/macro_parser.rs +++ b/src/librustc_expand/mbe/macro_parser.rs @@ -76,15 +76,11 @@ use TokenTreeOrTokenTreeSlice::*; use crate::mbe::{self, TokenTree}; -use rustc_ast::ptr::P; -use rustc_ast::token::{self, DocComment, Nonterminal, Token}; -use rustc_ast_pretty::pprust; -use rustc_parse::parser::{FollowedByType, Parser, PathStyle}; +use rustc_ast::token::{self, DocComment, Nonterminal, NonterminalKind, Token}; +use rustc_parse::parser::Parser; use rustc_session::parse::ParseSess; -use rustc_span::symbol::{kw, sym, Ident, MacroRulesNormalizedIdent, Symbol}; +use rustc_span::symbol::{kw, MacroRulesNormalizedIdent}; -use rustc_errors::PResult; -use rustc_span::Span; use smallvec::{smallvec, SmallVec}; use rustc_data_structures::fx::FxHashMap; @@ -576,7 +572,8 @@ fn inner_parse_loop<'root, 'tt>( TokenTree::MetaVarDecl(_, _, id) => { // Built-in nonterminals never start with these tokens, // so we can eliminate them from consideration. - if may_begin_with(token, id.name) { + let kind = NonterminalKind::from_symbol(id.name).unwrap(); + if Parser::nonterminal_may_begin_with(kind, token) { bb_items.push(item); } } @@ -738,8 +735,19 @@ pub(super) fn parse_tt(parser: &mut Cow<'_, Parser<'_>>, ms: &[TokenTree]) -> Na let mut item = bb_items.pop().unwrap(); if let TokenTree::MetaVarDecl(span, _, ident) = item.top_elts.get_tt(item.idx) { let match_cur = item.match_cur; - let nt = match parse_nt(parser.to_mut(), span, ident.name) { - Err(()) => return ErrorReported, + let kind = NonterminalKind::from_symbol(ident.name).unwrap(); + let nt = match parser.to_mut().parse_nonterminal(kind) { + Err(mut err) => { + err.span_label( + span, + format!( + "while parsing argument for this `{}` macro fragment", + ident.name + ), + ) + .emit(); + return ErrorReported; + } Ok(nt) => nt, }; item.push_match(match_cur, MatchedNonterminal(Lrc::new(nt))); @@ -754,178 +762,3 @@ pub(super) fn parse_tt(parser: &mut Cow<'_, Parser<'_>>, ms: &[TokenTree]) -> Na assert!(!cur_items.is_empty()); } } - -/// The token is an identifier, but not `_`. -/// We prohibit passing `_` to macros expecting `ident` for now. -fn get_macro_ident(token: &Token) -> Option<(Ident, bool)> { - token.ident().filter(|(ident, _)| ident.name != kw::Underscore) -} - -/// Checks whether a non-terminal may begin with a particular token. -/// -/// Returning `false` is a *stability guarantee* that such a matcher will *never* begin with that -/// token. Be conservative (return true) if not sure. -fn may_begin_with(token: &Token, name: Symbol) -> bool { - /// Checks whether the non-terminal may contain a single (non-keyword) identifier. - fn may_be_ident(nt: &token::Nonterminal) -> bool { - match *nt { - token::NtItem(_) | token::NtBlock(_) | token::NtVis(_) | token::NtLifetime(_) => false, - _ => true, - } - } - - match name { - sym::expr => { - token.can_begin_expr() - // This exception is here for backwards compatibility. - && !token.is_keyword(kw::Let) - } - sym::ty => token.can_begin_type(), - sym::ident => get_macro_ident(token).is_some(), - sym::literal => token.can_begin_literal_maybe_minus(), - sym::vis => match token.kind { - // The follow-set of :vis + "priv" keyword + interpolated - token::Comma | token::Ident(..) | token::Interpolated(..) => true, - _ => token.can_begin_type(), - }, - sym::block => match token.kind { - token::OpenDelim(token::Brace) => true, - token::Interpolated(ref nt) => match **nt { - token::NtItem(_) - | token::NtPat(_) - | token::NtTy(_) - | token::NtIdent(..) - | token::NtMeta(_) - | token::NtPath(_) - | token::NtVis(_) => false, // none of these may start with '{'. - _ => true, - }, - _ => false, - }, - sym::path | sym::meta => match token.kind { - token::ModSep | token::Ident(..) => true, - token::Interpolated(ref nt) => match **nt { - token::NtPath(_) | token::NtMeta(_) => true, - _ => may_be_ident(&nt), - }, - _ => false, - }, - sym::pat => match token.kind { - token::Ident(..) | // box, ref, mut, and other identifiers (can stricten) - token::OpenDelim(token::Paren) | // tuple pattern - token::OpenDelim(token::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) - token::Interpolated(ref nt) => may_be_ident(nt), - _ => false, - }, - sym::lifetime => match token.kind { - token::Lifetime(_) => true, - token::Interpolated(ref nt) => match **nt { - token::NtLifetime(_) | token::NtTT(_) => true, - _ => false, - }, - _ => false, - }, - _ => match token.kind { - token::CloseDelim(_) => false, - _ => true, - }, - } -} - -/// A call to the "black-box" parser to parse some Rust non-terminal. -/// -/// # Parameters -/// -/// - `p`: the "black-box" parser to use -/// - `sp`: the `Span` we want to parse -/// - `name`: the name of the metavar _matcher_ we want to match (e.g., `tt`, `ident`, `block`, -/// etc...) -/// -/// # Returns -/// -/// The parsed non-terminal. -fn parse_nt(p: &mut Parser<'_>, sp: Span, name: Symbol) -> Result { - // FIXME(Centril): Consider moving this to `parser.rs` to make - // the visibilities of the methods used below `pub(super)` at most. - if name == sym::tt { - return Ok(token::NtTT(p.parse_token_tree())); - } - parse_nt_inner(p, sp, name).map_err(|mut err| { - err.span_label(sp, format!("while parsing argument for this `{}` macro fragment", name)) - .emit() - }) -} - -fn parse_nt_inner<'a>(p: &mut Parser<'a>, sp: Span, name: Symbol) -> PResult<'a, Nonterminal> { - // Any `Nonterminal` which stores its tokens (currently `NtItem` and `NtExpr`) - // needs to have them force-captured here. - // A `macro_rules!` invocation may pass a captured item/expr to a proc-macro, - // which requires having captured tokens available. Since we cannot determine - // in advance whether or not a proc-macro will be (transitively) invoked, - // we always capture tokens for any `Nonterminal` which needs them. - Ok(match name { - sym::item => match p.collect_tokens(|this| this.parse_item())? { - (Some(mut item), tokens) => { - // If we captured tokens during parsing (due to outer attributes), - // use those. - if item.tokens.is_none() { - item.tokens = Some(tokens); - } - token::NtItem(item) - } - (None, _) => return Err(p.struct_span_err(p.token.span, "expected an item keyword")), - }, - sym::block => token::NtBlock(p.parse_block()?), - sym::stmt => match p.parse_stmt()? { - Some(s) => token::NtStmt(s), - None => return Err(p.struct_span_err(p.token.span, "expected a statement")), - }, - sym::pat => token::NtPat(p.parse_pat(None)?), - sym::expr => { - let (mut expr, tokens) = p.collect_tokens(|this| this.parse_expr())?; - // If we captured tokens during parsing (due to outer attributes), - // use those. - if expr.tokens.is_none() { - expr.tokens = Some(tokens); - } - token::NtExpr(expr) - } - sym::literal => token::NtLiteral(p.parse_literal_maybe_minus()?), - sym::ty => token::NtTy(p.parse_ty()?), - // this could be handled like a token, since it is one - sym::ident => { - if let Some((ident, is_raw)) = get_macro_ident(&p.token) { - p.bump(); - token::NtIdent(ident, is_raw) - } else { - let token_str = pprust::token_to_string(&p.token); - let msg = &format!("expected ident, found {}", &token_str); - return Err(p.struct_span_err(p.token.span, msg)); - } - } - sym::path => token::NtPath(p.parse_path(PathStyle::Type)?), - sym::meta => token::NtMeta(P(p.parse_attr_item()?)), - sym::vis => token::NtVis(p.parse_visibility(FollowedByType::Yes)?), - sym::lifetime => { - if p.check_lifetime() { - token::NtLifetime(p.expect_lifetime().ident) - } else { - let token_str = pprust::token_to_string(&p.token); - let msg = &format!("expected a lifetime, found `{}`", &token_str); - return Err(p.struct_span_err(p.token.span, msg)); - } - } - // this is not supposed to happen, since it has been checked - // when compiling the macro. - _ => p.span_bug(sp, "invalid fragment specifier"), - }) -} diff --git a/src/librustc_expand/mbe/macro_rules.rs b/src/librustc_expand/mbe/macro_rules.rs index 36b323df6979..657c72225928 100644 --- a/src/librustc_expand/mbe/macro_rules.rs +++ b/src/librustc_expand/mbe/macro_rules.rs @@ -9,7 +9,7 @@ use crate::mbe::macro_parser::{MatchedNonterminal, MatchedSeq}; use crate::mbe::transcribe::transcribe; use rustc_ast::ast; -use rustc_ast::token::{self, NtTT, Token, TokenKind::*}; +use rustc_ast::token::{self, NonterminalKind, NtTT, Token, TokenKind::*}; use rustc_ast::tokenstream::{DelimSpan, TokenStream}; use rustc_ast_pretty::pprust; use rustc_attr::{self as attr, TransparencyError}; @@ -1228,23 +1228,7 @@ fn is_legal_fragment_specifier( * for checking against feature gates. See past versions of * this function. */ - match frag_name { - sym::item - | sym::block - | sym::stmt - | sym::expr - | sym::pat - | sym::lifetime - | sym::path - | sym::ty - | sym::ident - | sym::meta - | sym::tt - | sym::vis - | sym::literal - | kw::Invalid => true, - _ => false, - } + NonterminalKind::from_symbol(frag_name).is_some() || frag_name == kw::Invalid } fn quoted_tt_to_string(tt: &mbe::TokenTree) -> String { diff --git a/src/librustc_parse/parser/expr.rs b/src/librustc_parse/parser/expr.rs index d06b172bc148..3aec300d86d4 100644 --- a/src/librustc_parse/parser/expr.rs +++ b/src/librustc_parse/parser/expr.rs @@ -1450,7 +1450,7 @@ impl<'a> Parser<'a> { /// Matches `'-' lit | lit` (cf. `ast_validation::AstValidator::check_expr_within_pat`). /// Keep this in sync with `Token::can_begin_literal_maybe_minus`. - pub fn parse_literal_maybe_minus(&mut self) -> PResult<'a, P> { + pub(super) fn parse_literal_maybe_minus(&mut self) -> PResult<'a, P> { maybe_whole_expr!(self); let lo = self.token.span; diff --git a/src/librustc_parse/parser/mod.rs b/src/librustc_parse/parser/mod.rs index 72866468b656..38d2bd9b7566 100644 --- a/src/librustc_parse/parser/mod.rs +++ b/src/librustc_parse/parser/mod.rs @@ -1,6 +1,7 @@ pub mod attr; mod expr; mod item; +mod nonterminal; mod pat; mod path; mod ty; @@ -10,6 +11,7 @@ mod generics; mod stmt; use diagnostics::Error; + use crate::lexer::UnmatchedBrace; use log::debug; @@ -958,7 +960,7 @@ impl<'a> Parser<'a> { } /// Parses a single token tree from the input. - pub fn parse_token_tree(&mut self) -> TokenTree { + pub(crate) fn parse_token_tree(&mut self) -> TokenTree { match self.token.kind { token::OpenDelim(..) => { let frame = mem::replace( @@ -1017,7 +1019,7 @@ impl<'a> Parser<'a> { /// If the following element can't be a tuple (i.e., it's a function definition), then /// it's not a tuple struct field), and the contents within the parentheses isn't valid, /// so emit a proper diagnostic. - pub fn parse_visibility(&mut self, fbt: FollowedByType) -> PResult<'a, Visibility> { + pub(crate) fn parse_visibility(&mut self, fbt: FollowedByType) -> PResult<'a, Visibility> { maybe_whole!(self, NtVis, |x| x); self.expected_tokens.push(TokenType::Keyword(kw::Crate)); diff --git a/src/librustc_parse/parser/nonterminal.rs b/src/librustc_parse/parser/nonterminal.rs new file mode 100644 index 000000000000..12139771bbfd --- /dev/null +++ b/src/librustc_parse/parser/nonterminal.rs @@ -0,0 +1,163 @@ +use rustc_ast::ptr::P; +use rustc_ast::token::{self, Nonterminal, NonterminalKind, Token}; +use rustc_ast_pretty::pprust; +use rustc_errors::PResult; +use rustc_span::symbol::{kw, Ident}; + +use crate::parser::{FollowedByType, Parser, PathStyle}; + +impl<'a> Parser<'a> { + /// Checks whether a non-terminal may begin with a particular token. + /// + /// Returning `false` is a *stability guarantee* that such a matcher will *never* begin with that + /// token. Be conservative (return true) if not sure. + pub fn nonterminal_may_begin_with(kind: NonterminalKind, token: &Token) -> bool { + /// Checks whether the non-terminal may contain a single (non-keyword) identifier. + fn may_be_ident(nt: &token::Nonterminal) -> bool { + match *nt { + token::NtItem(_) | token::NtBlock(_) | token::NtVis(_) | token::NtLifetime(_) => { + false + } + _ => true, + } + } + + match kind { + NonterminalKind::Expr => { + token.can_begin_expr() + // This exception is here for backwards compatibility. + && !token.is_keyword(kw::Let) + } + NonterminalKind::Ty => token.can_begin_type(), + NonterminalKind::Ident => get_macro_ident(token).is_some(), + NonterminalKind::Literal => token.can_begin_literal_maybe_minus(), + NonterminalKind::Vis => match token.kind { + // The follow-set of :vis + "priv" keyword + interpolated + token::Comma | token::Ident(..) | token::Interpolated(..) => true, + _ => token.can_begin_type(), + }, + NonterminalKind::Block => match token.kind { + token::OpenDelim(token::Brace) => true, + token::Interpolated(ref nt) => match **nt { + token::NtItem(_) + | token::NtPat(_) + | token::NtTy(_) + | token::NtIdent(..) + | token::NtMeta(_) + | token::NtPath(_) + | token::NtVis(_) => false, // none of these may start with '{'. + _ => true, + }, + _ => false, + }, + NonterminalKind::Path | NonterminalKind::Meta => match token.kind { + token::ModSep | token::Ident(..) => true, + token::Interpolated(ref nt) => match **nt { + token::NtPath(_) | token::NtMeta(_) => true, + _ => may_be_ident(&nt), + }, + _ => false, + }, + NonterminalKind::Pat => match token.kind { + token::Ident(..) | // box, ref, mut, and other identifiers (can stricten) + token::OpenDelim(token::Paren) | // tuple pattern + token::OpenDelim(token::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) + token::Interpolated(ref nt) => may_be_ident(nt), + _ => false, + }, + NonterminalKind::Lifetime => match token.kind { + token::Lifetime(_) => true, + token::Interpolated(ref nt) => match **nt { + token::NtLifetime(_) | token::NtTT(_) => true, + _ => false, + }, + _ => false, + }, + NonterminalKind::TT | NonterminalKind::Item | NonterminalKind::Stmt => match token.kind + { + token::CloseDelim(_) => false, + _ => true, + }, + } + } + + pub fn parse_nonterminal(&mut self, kind: NonterminalKind) -> PResult<'a, Nonterminal> { + // Any `Nonterminal` which stores its tokens (currently `NtItem` and `NtExpr`) + // needs to have them force-captured here. + // A `macro_rules!` invocation may pass a captured item/expr to a proc-macro, + // which requires having captured tokens available. Since we cannot determine + // in advance whether or not a proc-macro will be (transitively) invoked, + // we always capture tokens for any `Nonterminal` which needs them. + Ok(match kind { + NonterminalKind::Item => match self.collect_tokens(|this| this.parse_item())? { + (Some(mut item), tokens) => { + // If we captured tokens during parsing (due to outer attributes), + // use those. + if item.tokens.is_none() { + item.tokens = Some(tokens); + } + token::NtItem(item) + } + (None, _) => { + return Err(self.struct_span_err(self.token.span, "expected an item keyword")); + } + }, + NonterminalKind::Block => token::NtBlock(self.parse_block()?), + NonterminalKind::Stmt => match self.parse_stmt()? { + Some(s) => token::NtStmt(s), + None => return Err(self.struct_span_err(self.token.span, "expected a statement")), + }, + NonterminalKind::Pat => token::NtPat(self.parse_pat(None)?), + NonterminalKind::Expr => { + let (mut expr, tokens) = self.collect_tokens(|this| this.parse_expr())?; + // If we captured tokens during parsing (due to outer attributes), + // use those. + if expr.tokens.is_none() { + expr.tokens = Some(tokens); + } + token::NtExpr(expr) + } + NonterminalKind::Literal => token::NtLiteral(self.parse_literal_maybe_minus()?), + NonterminalKind::Ty => token::NtTy(self.parse_ty()?), + // this could be handled like a token, since it is one + NonterminalKind::Ident => { + if let Some((ident, is_raw)) = get_macro_ident(&self.token) { + self.bump(); + token::NtIdent(ident, is_raw) + } else { + let token_str = pprust::token_to_string(&self.token); + let msg = &format!("expected ident, found {}", &token_str); + return Err(self.struct_span_err(self.token.span, msg)); + } + } + NonterminalKind::Path => token::NtPath(self.parse_path(PathStyle::Type)?), + NonterminalKind::Meta => token::NtMeta(P(self.parse_attr_item()?)), + NonterminalKind::TT => token::NtTT(self.parse_token_tree()), + NonterminalKind::Vis => token::NtVis(self.parse_visibility(FollowedByType::Yes)?), + NonterminalKind::Lifetime => { + if self.check_lifetime() { + token::NtLifetime(self.expect_lifetime().ident) + } else { + let token_str = pprust::token_to_string(&self.token); + let msg = &format!("expected a lifetime, found `{}`", &token_str); + return Err(self.struct_span_err(self.token.span, msg)); + } + } + }) + } +} + +/// The token is an identifier, but not `_`. +/// We prohibit passing `_` to macros expecting `ident` for now. +fn get_macro_ident(token: &Token) -> Option<(Ident, bool)> { + token.ident().filter(|(ident, _)| ident.name != kw::Underscore) +} diff --git a/src/librustc_parse/parser/path.rs b/src/librustc_parse/parser/path.rs index 3dcefd362574..d4e44c54b127 100644 --- a/src/librustc_parse/parser/path.rs +++ b/src/librustc_parse/parser/path.rs @@ -125,7 +125,7 @@ impl<'a> Parser<'a> { /// `a::b::C::` (with disambiguator) /// `Fn(Args)` (without disambiguator) /// `Fn::(Args)` (with disambiguator) - pub fn parse_path(&mut self, style: PathStyle) -> PResult<'a, Path> { + pub(super) fn parse_path(&mut self, style: PathStyle) -> PResult<'a, Path> { maybe_whole!(self, NtPath, |path| { if style == PathStyle::Mod && path.segments.iter().any(|segment| segment.args.is_some()) { diff --git a/src/librustc_parse/parser/stmt.rs b/src/librustc_parse/parser/stmt.rs index d04920de47f0..5c3a5e998732 100644 --- a/src/librustc_parse/parser/stmt.rs +++ b/src/librustc_parse/parser/stmt.rs @@ -21,7 +21,7 @@ use std::mem; impl<'a> Parser<'a> { /// Parses a statement. This stops just before trailing semicolons on everything but items. /// e.g., a `StmtKind::Semi` parses to a `StmtKind::Expr`, leaving the trailing `;` unconsumed. - pub fn parse_stmt(&mut self) -> PResult<'a, Option> { + pub(super) fn parse_stmt(&mut self) -> PResult<'a, Option> { Ok(self.parse_stmt_without_recovery().unwrap_or_else(|mut e| { e.emit(); self.recover_stmt_(SemiColonMode::Break, BlockMode::Ignore); @@ -247,7 +247,7 @@ impl<'a> Parser<'a> { } /// Parses a block. No inner attributes are allowed. - pub fn parse_block(&mut self) -> PResult<'a, P> { + pub(super) fn parse_block(&mut self) -> PResult<'a, P> { let (attrs, block) = self.parse_inner_attrs_and_block()?; if let [.., last] = &*attrs { self.error_on_forbidden_inner_attr(last.span, DEFAULT_INNER_ATTR_FORBIDDEN); diff --git a/src/librustc_parse/parser/ty.rs b/src/librustc_parse/parser/ty.rs index a6015504a328..cd66b917f232 100644 --- a/src/librustc_parse/parser/ty.rs +++ b/src/librustc_parse/parser/ty.rs @@ -610,13 +610,13 @@ impl<'a> Parser<'a> { } } - pub fn check_lifetime(&mut self) -> bool { + pub(super) fn check_lifetime(&mut self) -> bool { self.expected_tokens.push(TokenType::Lifetime); self.token.is_lifetime() } /// Parses a single lifetime `'a` or panics. - pub fn expect_lifetime(&mut self) -> Lifetime { + pub(super) fn expect_lifetime(&mut self) -> Lifetime { if let Some(ident) = self.token.lifetime() { self.bump(); Lifetime { ident, id: ast::DUMMY_NODE_ID }