diff options
| author | bors <bors@rust-lang.org> | 2019-02-23 14:14:04 +0000 |
|---|---|---|
| committer | bors <bors@rust-lang.org> | 2019-02-23 14:14:04 +0000 |
| commit | 7f19f161f24c9a02ff8c3f73122d0b015039221f (patch) | |
| tree | e02114783d06bffa83bd348a8d5dec4bd0767721 /src/libsyntax/parse | |
| parent | bcfb5e8ac30c46ba512526d66f803756c124c7bb (diff) | |
| parent | a36d1b930108a33182d412832aa4f6cc1b37158f (diff) | |
Auto merge of #58669 - Centril:rollup, r=Centril
Rollup of 16 pull requests Successful merges: - #58100 (Transition librustdoc to Rust 2018) - #58122 (RangeInclusive internal iteration performance improvement.) - #58199 (Add better error message for partial move) - #58227 (Updated RELEASES.md for 1.33.0) - #58353 (Check the Self-type of inherent associated constants) - #58453 (SGX target: fix panic = abort) - #58476 (Remove `LazyTokenStream`.) - #58526 (Special suggestion for illegal unicode curly quote pairs) - #58595 (Turn duration consts into associated consts) - #58609 (Allow Self::Module to be mutated.) - #58628 (Optimise vec![false; N] to zero-alloc) - #58643 (Don't generate minification variables if minification disabled) - #58648 (Update tests to account for cross-platform testing and miri.) - #58654 (Do not underflow after resetting unmatched braces count) - #58658 (Add expected/provided byte alignments to validation error message) - #58667 (Reduce Miri-related Code Repetition `like (n << amt) >> amt`) Failed merges: r? @ghost
Diffstat (limited to 'src/libsyntax/parse')
| -rw-r--r-- | src/libsyntax/parse/attr.rs | 4 | ||||
| -rw-r--r-- | src/libsyntax/parse/lexer/mod.rs | 22 | ||||
| -rw-r--r-- | src/libsyntax/parse/lexer/unicode_chars.rs | 31 | ||||
| -rw-r--r-- | src/libsyntax/parse/parser.rs | 28 | ||||
| -rw-r--r-- | src/libsyntax/parse/token.rs | 248 |
5 files changed, 154 insertions, 179 deletions
diff --git a/src/libsyntax/parse/attr.rs b/src/libsyntax/parse/attr.rs index b36ca0574cb..9020c8c6a2d 100644 --- a/src/libsyntax/parse/attr.rs +++ b/src/libsyntax/parse/attr.rs @@ -141,7 +141,7 @@ impl<'a> Parser<'a> { /// The delimiters or `=` are still put into the resulting token stream. crate fn parse_meta_item_unrestricted(&mut self) -> PResult<'a, (ast::Path, TokenStream)> { let meta = match self.token { - token::Interpolated(ref nt) => match nt.0 { + token::Interpolated(ref nt) => match **nt { Nonterminal::NtMeta(ref meta) => Some(meta.clone()), _ => None, }, @@ -227,7 +227,7 @@ impl<'a> Parser<'a> { /// meta_item_inner : (meta_item | UNSUFFIXED_LIT) (',' meta_item_inner)? ; pub fn parse_meta_item(&mut self) -> PResult<'a, ast::MetaItem> { let nt_meta = match self.token { - token::Interpolated(ref nt) => match nt.0 { + token::Interpolated(ref nt) => match **nt { token::NtMeta(ref e) => Some(e.clone()), _ => None, }, diff --git a/src/libsyntax/parse/lexer/mod.rs b/src/libsyntax/parse/lexer/mod.rs index babe0eef20f..a7cde5fbb92 100644 --- a/src/libsyntax/parse/lexer/mod.rs +++ b/src/libsyntax/parse/lexer/mod.rs @@ -125,6 +125,28 @@ impl<'a> StringReader<'a> { Ok(ret_val) } + /// Immutably extract string if found at current position with given delimiters + pub fn peek_delimited(&self, from_ch: char, to_ch: char) -> Option<String> { + let mut pos = self.pos; + let mut idx = self.src_index(pos); + let mut ch = char_at(&self.src, idx); + if ch != from_ch { + return None; + } + pos = pos + Pos::from_usize(ch.len_utf8()); + let start_pos = pos; + idx = self.src_index(pos); + while idx < self.end_src_index { + ch = char_at(&self.src, idx); + if ch == to_ch { + return Some(self.src[self.src_index(start_pos)..self.src_index(pos)].to_string()); + } + pos = pos + Pos::from_usize(ch.len_utf8()); + idx = self.src_index(pos); + } + return None; + } + fn try_real_token(&mut self) -> Result<TokenAndSpan, ()> { let mut t = self.try_next_token()?; loop { diff --git a/src/libsyntax/parse/lexer/unicode_chars.rs b/src/libsyntax/parse/lexer/unicode_chars.rs index 7da4284c0e4..94ce6297fbe 100644 --- a/src/libsyntax/parse/lexer/unicode_chars.rs +++ b/src/libsyntax/parse/lexer/unicode_chars.rs @@ -1,7 +1,7 @@ // Characters and their corresponding confusables were collected from // http://www.unicode.org/Public/security/10.0.0/confusables.txt -use syntax_pos::{Span, NO_EXPANSION}; +use syntax_pos::{Span, Pos, NO_EXPANSION}; use errors::{Applicability, DiagnosticBuilder}; use super::StringReader; @@ -333,14 +333,27 @@ crate fn check_for_substitution<'a>(reader: &StringReader<'a>, let span = Span::new(reader.pos, reader.next_pos, NO_EXPANSION); match ASCII_ARRAY.iter().find(|&&(c, _)| c == ascii_char) { Some(&(ascii_char, ascii_name)) => { - let msg = - format!("Unicode character '{}' ({}) looks like '{}' ({}), but it is not", - ch, u_name, ascii_char, ascii_name); - err.span_suggestion( - span, - &msg, - ascii_char.to_string(), - Applicability::MaybeIncorrect); + // special help suggestion for "directed" double quotes + if let Some(s) = reader.peek_delimited('“', '”') { + let msg = format!("Unicode characters '“' (Left Double Quotation Mark) and \ + '”' (Right Double Quotation Mark) look like '{}' ({}), but are not", + ascii_char, ascii_name); + err.span_suggestion( + Span::new(reader.pos, reader.next_pos + Pos::from_usize(s.len()) + + Pos::from_usize('”'.len_utf8()), NO_EXPANSION), + &msg, + format!("\"{}\"", s), + Applicability::MaybeIncorrect); + } else { + let msg = + format!("Unicode character '{}' ({}) looks like '{}' ({}), but it is not", + ch, u_name, ascii_char, ascii_name); + err.span_suggestion( + span, + &msg, + ascii_char.to_string(), + Applicability::MaybeIncorrect); + } true }, None => { diff --git a/src/libsyntax/parse/parser.rs b/src/libsyntax/parse/parser.rs index e6a48912c48..5a753e1f8c8 100644 --- a/src/libsyntax/parse/parser.rs +++ b/src/libsyntax/parse/parser.rs @@ -119,7 +119,7 @@ enum BlockMode { macro_rules! maybe_whole_expr { ($p:expr) => { if let token::Interpolated(nt) = $p.token.clone() { - match nt.0 { + match *nt { token::NtExpr(ref e) | token::NtLiteral(ref e) => { $p.bump(); return Ok((*e).clone()); @@ -146,7 +146,7 @@ macro_rules! maybe_whole_expr { macro_rules! maybe_whole { ($p:expr, $constructor:ident, |$x:ident| $e:expr) => { if let token::Interpolated(nt) = $p.token.clone() { - if let token::$constructor($x) = nt.0.clone() { + if let token::$constructor($x) = (*nt).clone() { $p.bump(); return Ok($e); } @@ -1184,8 +1184,10 @@ impl<'a> Parser<'a> { match ate { Some(_) => { // See doc comment for `unmatched_angle_bracket_count`. - self.unmatched_angle_bracket_count -= 1; - debug!("expect_gt: (decrement) count={:?}", self.unmatched_angle_bracket_count); + if self.unmatched_angle_bracket_count > 0 { + self.unmatched_angle_bracket_count -= 1; + debug!("expect_gt: (decrement) count={:?}", self.unmatched_angle_bracket_count); + } Ok(()) }, @@ -1570,7 +1572,7 @@ impl<'a> Parser<'a> { Some(body) } token::Interpolated(ref nt) => { - match &nt.0 { + match **nt { token::NtBlock(..) => { *at_end = true; let (inner_attrs, body) = self.parse_inner_attrs_and_block()?; @@ -1913,7 +1915,7 @@ impl<'a> Parser<'a> { fn is_named_argument(&mut self) -> bool { let offset = match self.token { - token::Interpolated(ref nt) => match nt.0 { + token::Interpolated(ref nt) => match **nt { token::NtPat(..) => return self.look_ahead(1, |t| t == &token::Colon), _ => 0, } @@ -2099,7 +2101,7 @@ impl<'a> Parser<'a> { /// Matches `token_lit = LIT_INTEGER | ...`. fn parse_lit_token(&mut self) -> PResult<'a, LitKind> { let out = match self.token { - token::Interpolated(ref nt) => match nt.0 { + token::Interpolated(ref nt) => match **nt { token::NtExpr(ref v) | token::NtLiteral(ref v) => match v.node { ExprKind::Lit(ref lit) => { lit.node.clone() } _ => { return self.unexpected_last(&self.token); } @@ -2248,8 +2250,10 @@ impl<'a> Parser<'a> { // See doc comment for `unmatched_angle_bracket_count`. self.expect(&token::Gt)?; - self.unmatched_angle_bracket_count -= 1; - debug!("parse_qpath: (decrement) count={:?}", self.unmatched_angle_bracket_count); + if self.unmatched_angle_bracket_count > 0 { + self.unmatched_angle_bracket_count -= 1; + debug!("parse_qpath: (decrement) count={:?}", self.unmatched_angle_bracket_count); + } self.expect(&token::ModSep)?; @@ -2299,7 +2303,7 @@ impl<'a> Parser<'a> { /// attributes. pub fn parse_path_allowing_meta(&mut self, style: PathStyle) -> PResult<'a, ast::Path> { let meta_ident = match self.token { - token::Interpolated(ref nt) => match nt.0 { + token::Interpolated(ref nt) => match **nt { token::NtMeta(ref meta) => match meta.node { ast::MetaItemKind::Word => Some(meta.ident.clone()), _ => None, @@ -3271,7 +3275,7 @@ impl<'a> Parser<'a> { self.meta_var_span = Some(self.span); // Interpolated identifier and lifetime tokens are replaced with usual identifier // and lifetime tokens, so the former are never encountered during normal parsing. - match nt.0 { + match **nt { token::NtIdent(ident, is_raw) => (token::Ident(ident, is_raw), ident.span), token::NtLifetime(ident) => (token::Lifetime(ident), ident.span), _ => return, @@ -3403,7 +3407,7 @@ impl<'a> Parser<'a> { // can't continue an expression after an ident token::Ident(ident, is_raw) => token::ident_can_begin_expr(ident, is_raw), token::Literal(..) | token::Pound => true, - token::Interpolated(ref nt) => match nt.0 { + token::Interpolated(ref nt) => match **nt { token::NtIdent(..) | token::NtExpr(..) | token::NtBlock(..) | token::NtPath(..) => true, _ => false, diff --git a/src/libsyntax/parse/token.rs b/src/libsyntax/parse/token.rs index ff7f3e0bfae..eec422d6266 100644 --- a/src/libsyntax/parse/token.rs +++ b/src/libsyntax/parse/token.rs @@ -13,16 +13,15 @@ use crate::syntax::parse::parse_stream_from_source_str; use crate::syntax::parse::parser::emit_unclosed_delims; use crate::tokenstream::{self, DelimSpan, TokenStream, TokenTree}; -use serialize::{Decodable, Decoder, Encodable, Encoder}; use syntax_pos::symbol::{self, Symbol}; use syntax_pos::{self, Span, FileName}; use log::info; -use std::{cmp, fmt}; +use std::fmt; use std::mem; #[cfg(target_arch = "x86_64")] use rustc_data_structures::static_assert; -use rustc_data_structures::sync::{Lrc, Lock}; +use rustc_data_structures::sync::Lrc; #[derive(Clone, PartialEq, RustcEncodable, RustcDecodable, Hash, Debug, Copy)] pub enum BinOpToken { @@ -87,7 +86,7 @@ impl Lit { } } - // See comments in `interpolated_to_tokenstream` for why we care about + // See comments in `Nonterminal::to_tokenstream` for why we care about // *probably* equal here rather than actual equality fn probably_equal_for_proc_macro(&self, other: &Lit) -> bool { mem::discriminant(self) == mem::discriminant(other) @@ -184,9 +183,8 @@ pub enum Token { Ident(ast::Ident, /* is_raw */ bool), Lifetime(ast::Ident), - // The `LazyTokenStream` is a pure function of the `Nonterminal`, - // and so the `LazyTokenStream` can be ignored by Eq, Hash, etc. - Interpolated(Lrc<(Nonterminal, LazyTokenStream)>), + Interpolated(Lrc<Nonterminal>), + // Can be expanded into several tokens. /// A doc comment. DocComment(ast::Name), @@ -209,10 +207,6 @@ pub enum Token { static_assert!(MEM_SIZE_OF_STATEMENT: mem::size_of::<Token>() == 16); impl Token { - pub fn interpolated(nt: Nonterminal) -> Token { - Token::Interpolated(Lrc::new((nt, LazyTokenStream::new()))) - } - /// Recovers a `Token` from an `ast::Ident`. This creates a raw identifier if necessary. pub fn from_ast_ident(ident: ast::Ident) -> Token { Ident(ident, ident.is_raw_guess()) @@ -244,7 +238,7 @@ impl Token { ModSep | // global path Lifetime(..) | // labeled loop Pound => true, // expression attributes - Interpolated(ref nt) => match nt.0 { + Interpolated(ref nt) => match **nt { NtLiteral(..) | NtIdent(..) | NtExpr(..) | @@ -272,7 +266,7 @@ impl Token { Lifetime(..) | // lifetime bound in trait object Lt | BinOp(Shl) | // associated path ModSep => true, // global path - Interpolated(ref nt) => match nt.0 { + Interpolated(ref nt) => match **nt { NtIdent(..) | NtTy(..) | NtPath(..) | NtLifetime(..) => true, _ => false, }, @@ -284,7 +278,7 @@ impl Token { pub fn can_begin_const_arg(&self) -> bool { match self { OpenDelim(Brace) => true, - Interpolated(ref nt) => match nt.0 { + Interpolated(ref nt) => match **nt { NtExpr(..) => true, NtBlock(..) => true, NtLiteral(..) => true, @@ -316,7 +310,7 @@ impl Token { BinOp(Minus) => true, Ident(ident, false) if ident.name == keywords::True.name() => true, Ident(ident, false) if ident.name == keywords::False.name() => true, - Interpolated(ref nt) => match nt.0 { + Interpolated(ref nt) => match **nt { NtLiteral(..) => true, _ => false, }, @@ -328,7 +322,7 @@ impl Token { pub fn ident(&self) -> Option<(ast::Ident, /* is_raw */ bool)> { match *self { Ident(ident, is_raw) => Some((ident, is_raw)), - Interpolated(ref nt) => match nt.0 { + Interpolated(ref nt) => match **nt { NtIdent(ident, is_raw) => Some((ident, is_raw)), _ => None, }, @@ -339,7 +333,7 @@ impl Token { pub fn lifetime(&self) -> Option<ast::Ident> { match *self { Lifetime(ident) => Some(ident), - Interpolated(ref nt) => match nt.0 { + Interpolated(ref nt) => match **nt { NtLifetime(ident) => Some(ident), _ => None, }, @@ -367,7 +361,7 @@ impl Token { /// Returns `true` if the token is an interpolated path. fn is_path(&self) -> bool { if let Interpolated(ref nt) = *self { - if let NtPath(..) = nt.0 { + if let NtPath(..) = **nt { return true; } } @@ -508,98 +502,7 @@ impl Token { } } - pub fn interpolated_to_tokenstream(&self, sess: &ParseSess, span: Span) - -> TokenStream - { - let nt = match *self { - Token::Interpolated(ref nt) => nt, - _ => panic!("only works on interpolated tokens"), - }; - - // An `Interpolated` token means that we have a `Nonterminal` - // which is often a parsed AST item. At this point we now need - // to convert the parsed AST to an actual token stream, e.g. - // un-parse it basically. - // - // Unfortunately there's not really a great way to do that in a - // guaranteed lossless fashion right now. The fallback here is - // to just stringify the AST node and reparse it, but this loses - // all span information. - // - // As a result, some AST nodes are annotated with the token - // stream they came from. Here we attempt to extract these - // lossless token streams before we fall back to the - // stringification. - let mut tokens = None; - - match nt.0 { - Nonterminal::NtItem(ref item) => { - tokens = prepend_attrs(sess, &item.attrs, item.tokens.as_ref(), span); - } - Nonterminal::NtTraitItem(ref item) => { - tokens = prepend_attrs(sess, &item.attrs, item.tokens.as_ref(), span); - } - Nonterminal::NtImplItem(ref item) => { - tokens = prepend_attrs(sess, &item.attrs, item.tokens.as_ref(), span); - } - Nonterminal::NtIdent(ident, is_raw) => { - let token = Token::Ident(ident, is_raw); - tokens = Some(TokenTree::Token(ident.span, token).into()); - } - Nonterminal::NtLifetime(ident) => { - let token = Token::Lifetime(ident); - tokens = Some(TokenTree::Token(ident.span, token).into()); - } - Nonterminal::NtTT(ref tt) => { - tokens = Some(tt.clone().into()); - } - _ => {} - } - - let tokens_for_real = nt.1.force(|| { - // FIXME(#43081): Avoid this pretty-print + reparse hack - let source = pprust::token_to_string(self); - let filename = FileName::macro_expansion_source_code(&source); - let (tokens, errors) = parse_stream_from_source_str( - filename, source, sess, Some(span)); - emit_unclosed_delims(&errors, &sess.span_diagnostic); - tokens - }); - - // During early phases of the compiler the AST could get modified - // directly (e.g., attributes added or removed) and the internal cache - // of tokens my not be invalidated or updated. Consequently if the - // "lossless" token stream disagrees with our actual stringification - // (which has historically been much more battle-tested) then we go - // with the lossy stream anyway (losing span information). - // - // Note that the comparison isn't `==` here to avoid comparing spans, - // but it *also* is a "probable" equality which is a pretty weird - // definition. We mostly want to catch actual changes to the AST - // like a `#[cfg]` being processed or some weird `macro_rules!` - // expansion. - // - // What we *don't* want to catch is the fact that a user-defined - // literal like `0xf` is stringified as `15`, causing the cached token - // stream to not be literal `==` token-wise (ignoring spans) to the - // token stream we got from stringification. - // - // Instead the "probably equal" check here is "does each token - // recursively have the same discriminant?" We basically don't look at - // the token values here and assume that such fine grained token stream - // modifications, including adding/removing typically non-semantic - // tokens such as extra braces and commas, don't happen. - if let Some(tokens) = tokens { - if tokens.probably_equal_for_proc_macro(&tokens_for_real) { - return tokens - } - info!("cached tokens found, but they're not \"probably equal\", \ - going with stringified version"); - } - return tokens_for_real - } - - // See comments in `interpolated_to_tokenstream` for why we care about + // See comments in `Nonterminal::to_tokenstream` for why we care about // *probably* equal here rather than actual equality crate fn probably_equal_for_proc_macro(&self, other: &Token) -> bool { if mem::discriminant(self) != mem::discriminant(other) { @@ -731,61 +634,94 @@ impl fmt::Debug for Nonterminal { } } -crate fn is_op(tok: &Token) -> bool { - match *tok { - OpenDelim(..) | CloseDelim(..) | Literal(..) | DocComment(..) | - Ident(..) | Lifetime(..) | Interpolated(..) | - Whitespace | Comment | Shebang(..) | Eof => false, - _ => true, - } -} - -#[derive(Clone)] -pub struct LazyTokenStream(Lock<Option<TokenStream>>); - -impl cmp::Eq for LazyTokenStream {} -impl PartialEq for LazyTokenStream { - fn eq(&self, _other: &LazyTokenStream) -> bool { - true - } -} - -impl fmt::Debug for LazyTokenStream { - fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { - fmt::Debug::fmt(&self.clone().0.into_inner(), f) - } -} +impl Nonterminal { + pub fn to_tokenstream(&self, sess: &ParseSess, span: Span) -> TokenStream { + // A `Nonterminal` is often a parsed AST item. At this point we now + // need to convert the parsed AST to an actual token stream, e.g. + // un-parse it basically. + // + // Unfortunately there's not really a great way to do that in a + // guaranteed lossless fashion right now. The fallback here is to just + // stringify the AST node and reparse it, but this loses all span + // information. + // + // As a result, some AST nodes are annotated with the token stream they + // came from. Here we attempt to extract these lossless token streams + // before we fall back to the stringification. + let tokens = match *self { + Nonterminal::NtItem(ref item) => { + prepend_attrs(sess, &item.attrs, item.tokens.as_ref(), span) + } + Nonterminal::NtTraitItem(ref item) => { + prepend_attrs(sess, &item.attrs, item.tokens.as_ref(), span) + } + Nonterminal::NtImplItem(ref item) => { + prepend_attrs(sess, &item.attrs, item.tokens.as_ref(), span) + } + Nonterminal::NtIdent(ident, is_raw) => { + let token = Token::Ident(ident, is_raw); + Some(TokenTree::Token(ident.span, token).into()) + } + Nonterminal::NtLifetime(ident) => { + let token = Token::Lifetime(ident); + Some(TokenTree::Token(ident.span, token).into()) + } + Nonterminal::NtTT(ref tt) => { + Some(tt.clone().into()) + } + _ => None, + }; -impl LazyTokenStream { - pub fn new() -> Self { - LazyTokenStream(Lock::new(None)) - } + // FIXME(#43081): Avoid this pretty-print + reparse hack + let source = pprust::nonterminal_to_string(self); + let filename = FileName::macro_expansion_source_code(&source); + let (tokens_for_real, errors) = + parse_stream_from_source_str(filename, source, sess, Some(span)); + emit_unclosed_delims(&errors, &sess.span_diagnostic); - fn force<F: FnOnce() -> TokenStream>(&self, f: F) -> TokenStream { - let mut opt_stream = self.0.lock(); - if opt_stream.is_none() { - *opt_stream = Some(f()); + // During early phases of the compiler the AST could get modified + // directly (e.g., attributes added or removed) and the internal cache + // of tokens my not be invalidated or updated. Consequently if the + // "lossless" token stream disagrees with our actual stringification + // (which has historically been much more battle-tested) then we go + // with the lossy stream anyway (losing span information). + // + // Note that the comparison isn't `==` here to avoid comparing spans, + // but it *also* is a "probable" equality which is a pretty weird + // definition. We mostly want to catch actual changes to the AST + // like a `#[cfg]` being processed or some weird `macro_rules!` + // expansion. + // + // What we *don't* want to catch is the fact that a user-defined + // literal like `0xf` is stringified as `15`, causing the cached token + // stream to not be literal `==` token-wise (ignoring spans) to the + // token stream we got from stringification. + // + // Instead the "probably equal" check here is "does each token + // recursively have the same discriminant?" We basically don't look at + // the token values here and assume that such fine grained token stream + // modifications, including adding/removing typically non-semantic + // tokens such as extra braces and commas, don't happen. + if let Some(tokens) = tokens { + if tokens.probably_equal_for_proc_macro(&tokens_for_real) { + return tokens + } + info!("cached tokens found, but they're not \"probably equal\", \ + going with stringified version"); } - opt_stream.clone().unwrap() - } -} - -impl Encodable for LazyTokenStream { - fn encode<S: Encoder>(&self, _: &mut S) -> Result<(), S::Error> { - Ok(()) + return tokens_for_real } } -impl Decodable for LazyTokenStream { - fn decode<D: Decoder>(_: &mut D) -> Result<LazyTokenStream, D::Error> { - Ok(LazyTokenStream::new()) +crate fn is_op(tok: &Token) -> bool { + match *tok { + OpenDelim(..) | CloseDelim(..) | Literal(..) | DocComment(..) | + Ident(..) | Lifetime(..) | Interpolated(..) | + Whitespace | Comment | Shebang(..) | Eof => false, + _ => true, } } -impl ::std::hash::Hash for LazyTokenStream { - fn hash<H: ::std::hash::Hasher>(&self, _hasher: &mut H) {} -} - fn prepend_attrs(sess: &ParseSess, attrs: &[ast::Attribute], tokens: Option<&tokenstream::TokenStream>, |
