diff options
Diffstat (limited to 'compiler/rustc_parse/src')
| -rw-r--r-- | compiler/rustc_parse/src/lexer/mod.rs | 41 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/lib.rs | 37 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/parser/expr.rs | 5 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/parser/mod.rs | 2 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/parser/pat.rs | 2 |
5 files changed, 62 insertions, 25 deletions
diff --git a/compiler/rustc_parse/src/lexer/mod.rs b/compiler/rustc_parse/src/lexer/mod.rs index 9792240a548..f5f081efc49 100644 --- a/compiler/rustc_parse/src/lexer/mod.rs +++ b/compiler/rustc_parse/src/lexer/mod.rs @@ -6,7 +6,7 @@ use rustc_ast::util::unicode::{TEXT_FLOW_CONTROL_CHARS, contains_text_flow_contr use rustc_errors::codes::*; use rustc_errors::{Applicability, Diag, DiagCtxtHandle, StashKey}; use rustc_lexer::{ - Base, Cursor, DocStyle, FrontmatterAllowed, LiteralKind, RawStrError, is_whitespace, + Base, Cursor, DocStyle, FrontmatterAllowed, LiteralKind, RawStrError, is_horizontal_whitespace, }; use rustc_literal_escaper::{EscapeError, Mode, check_for_errors}; use rustc_session::lint::BuiltinLintDiag; @@ -44,19 +44,44 @@ pub(crate) struct UnmatchedDelim { pub candidate_span: Option<Span>, } +/// Which tokens should be stripped before lexing the tokens. +pub(crate) enum StripTokens { + /// Strip both shebang and frontmatter. + ShebangAndFrontmatter, + /// Strip the shebang but not frontmatter. + /// + /// That means that char sequences looking like frontmatter are simply + /// interpreted as regular Rust lexemes. + Shebang, + /// Strip nothing. + /// + /// In other words, char sequences looking like a shebang or frontmatter + /// are simply interpreted as regular Rust lexemes. + Nothing, +} + pub(crate) fn lex_token_trees<'psess, 'src>( psess: &'psess ParseSess, mut src: &'src str, mut start_pos: BytePos, override_span: Option<Span>, - frontmatter_allowed: FrontmatterAllowed, + strip_tokens: StripTokens, ) -> Result<TokenStream, Vec<Diag<'psess>>> { - // Skip `#!`, if present. - if let Some(shebang_len) = rustc_lexer::strip_shebang(src) { - src = &src[shebang_len..]; - start_pos = start_pos + BytePos::from_usize(shebang_len); + match strip_tokens { + StripTokens::Shebang | StripTokens::ShebangAndFrontmatter => { + if let Some(shebang_len) = rustc_lexer::strip_shebang(src) { + src = &src[shebang_len..]; + start_pos = start_pos + BytePos::from_usize(shebang_len); + } + } + StripTokens::Nothing => {} } + let frontmatter_allowed = match strip_tokens { + StripTokens::ShebangAndFrontmatter => FrontmatterAllowed::Yes, + StripTokens::Shebang | StripTokens::Nothing => FrontmatterAllowed::No, + }; + let cursor = Cursor::new(src, frontmatter_allowed); let mut lexer = Lexer { psess, @@ -597,7 +622,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> { let last_line_start = within.rfind('\n').map_or(0, |i| i + 1); let last_line = &within[last_line_start..]; - let last_line_trimmed = last_line.trim_start_matches(is_whitespace); + let last_line_trimmed = last_line.trim_start_matches(is_horizontal_whitespace); let last_line_start_pos = frontmatter_opening_end_pos + BytePos(last_line_start as u32); let frontmatter_span = self.mk_sp(frontmatter_opening_pos, self.pos); @@ -640,7 +665,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> { }); } - if !rest.trim_matches(is_whitespace).is_empty() { + if !rest.trim_matches(is_horizontal_whitespace).is_empty() { let span = self.mk_sp(last_line_start_pos, self.pos); self.dcx().emit_err(errors::FrontmatterExtraCharactersAfterClose { span }); } diff --git a/compiler/rustc_parse/src/lib.rs b/compiler/rustc_parse/src/lib.rs index 197333d942d..d8792d7af4c 100644 --- a/compiler/rustc_parse/src/lib.rs +++ b/compiler/rustc_parse/src/lib.rs @@ -21,7 +21,6 @@ use rustc_ast::tokenstream::{DelimSpan, TokenStream}; use rustc_ast::{AttrItem, Attribute, MetaItemInner, token}; use rustc_ast_pretty::pprust; use rustc_errors::{Diag, EmissionGuarantee, FatalError, PResult, pluralize}; -use rustc_lexer::FrontmatterAllowed; use rustc_session::parse::ParseSess; use rustc_span::source_map::SourceMap; use rustc_span::{FileName, SourceFile, Span}; @@ -34,6 +33,8 @@ pub mod parser; use parser::Parser; use rustc_ast::token::Delimiter; +use crate::lexer::StripTokens; + pub mod lexer; mod errors; @@ -62,7 +63,20 @@ pub fn new_parser_from_source_str( source: String, ) -> Result<Parser<'_>, Vec<Diag<'_>>> { let source_file = psess.source_map().new_source_file(name, source); - new_parser_from_source_file(psess, source_file) + new_parser_from_source_file(psess, source_file, StripTokens::ShebangAndFrontmatter) +} + +/// Creates a new parser from a simple (no shebang, no frontmatter) source string. +/// +/// On failure, the errors must be consumed via `unwrap_or_emit_fatal`, `emit`, `cancel`, +/// etc., otherwise a panic will occur when they are dropped. +pub fn new_parser_from_simple_source_str( + psess: &ParseSess, + name: FileName, + source: String, +) -> Result<Parser<'_>, Vec<Diag<'_>>> { + let source_file = psess.source_map().new_source_file(name, source); + new_parser_from_source_file(psess, source_file, StripTokens::Nothing) } /// Creates a new parser from a filename. On failure, the errors must be consumed via @@ -96,7 +110,7 @@ pub fn new_parser_from_file<'a>( } err.emit(); }); - new_parser_from_source_file(psess, source_file) + new_parser_from_source_file(psess, source_file, StripTokens::ShebangAndFrontmatter) } pub fn utf8_error<E: EmissionGuarantee>( @@ -147,9 +161,10 @@ pub fn utf8_error<E: EmissionGuarantee>( fn new_parser_from_source_file( psess: &ParseSess, source_file: Arc<SourceFile>, + strip_tokens: StripTokens, ) -> Result<Parser<'_>, Vec<Diag<'_>>> { let end_pos = source_file.end_position(); - let stream = source_file_to_stream(psess, source_file, None, FrontmatterAllowed::Yes)?; + let stream = source_file_to_stream(psess, source_file, None, strip_tokens)?; let mut parser = Parser::new(psess, stream, None); if parser.token == token::Eof { parser.token.span = Span::new(end_pos, end_pos, parser.token.span.ctxt(), None); @@ -165,8 +180,8 @@ pub fn source_str_to_stream( ) -> Result<TokenStream, Vec<Diag<'_>>> { let source_file = psess.source_map().new_source_file(name, source); // used mainly for `proc_macro` and the likes, not for our parsing purposes, so don't parse - // frontmatters as frontmatters. - source_file_to_stream(psess, source_file, override_span, FrontmatterAllowed::No) + // frontmatters as frontmatters, but for compatibility reason still strip the shebang + source_file_to_stream(psess, source_file, override_span, StripTokens::Shebang) } /// Given a source file, produces a sequence of token trees. Returns any buffered errors from @@ -175,7 +190,7 @@ fn source_file_to_stream<'psess>( psess: &'psess ParseSess, source_file: Arc<SourceFile>, override_span: Option<Span>, - frontmatter_allowed: FrontmatterAllowed, + strip_tokens: StripTokens, ) -> Result<TokenStream, Vec<Diag<'psess>>> { let src = source_file.src.as_ref().unwrap_or_else(|| { psess.dcx().bug(format!( @@ -184,13 +199,7 @@ fn source_file_to_stream<'psess>( )); }); - lexer::lex_token_trees( - psess, - src.as_str(), - source_file.start_pos, - override_span, - frontmatter_allowed, - ) + lexer::lex_token_trees(psess, src.as_str(), source_file.start_pos, override_span, strip_tokens) } /// Runs the given subparser `f` on the tokens of the given `attr`'s item. diff --git a/compiler/rustc_parse/src/parser/expr.rs b/compiler/rustc_parse/src/parser/expr.rs index 7de4f6efd0b..4c02547357e 100644 --- a/compiler/rustc_parse/src/parser/expr.rs +++ b/compiler/rustc_parse/src/parser/expr.rs @@ -2742,7 +2742,7 @@ impl<'a> Parser<'a> { /// The specified `edition` in `let_chains_policy` should be that of the whole `if` construct, /// i.e. the same span we use to later decide whether the drop behaviour should be that of /// edition `..=2021` or that of `2024..`. - // Public because it is used in rustfmt forks such as https://github.com/tucant/rustfmt/blob/30c83df9e1db10007bdd16dafce8a86b404329b2/src/parse/macros/html.rs#L57 for custom if expressions. + // Public to use it for custom `if` expressions in rustfmt forks like https://github.com/tucant/rustfmt pub fn parse_expr_cond( &mut self, let_chains_policy: LetChainsPolicy, @@ -2910,7 +2910,8 @@ impl<'a> Parser<'a> { } } - fn parse_for_head(&mut self) -> PResult<'a, (Box<Pat>, Box<Expr>)> { + // Public to use it for custom `for` expressions in rustfmt forks like https://github.com/tucant/rustfmt + pub fn parse_for_head(&mut self) -> PResult<'a, (Box<Pat>, Box<Expr>)> { let begin_paren = if self.token == token::OpenParen { // Record whether we are about to parse `for (`. // This is used below for recovery in case of `for ( $stuff ) $block` diff --git a/compiler/rustc_parse/src/parser/mod.rs b/compiler/rustc_parse/src/parser/mod.rs index 15598f32429..3bbca622975 100644 --- a/compiler/rustc_parse/src/parser/mod.rs +++ b/compiler/rustc_parse/src/parser/mod.rs @@ -22,6 +22,8 @@ use std::{fmt, mem, slice}; use attr_wrapper::{AttrWrapper, UsePreAttrPos}; pub use diagnostics::AttemptLocalParseRecovery; pub(crate) use expr::ForbiddenLetReason; +// Public to use it for custom `if` expressions in rustfmt forks like https://github.com/tucant/rustfmt +pub use expr::LetChainsPolicy; pub(crate) use item::{FnContext, FnParseMode}; pub use pat::{CommaRecoveryMode, RecoverColon, RecoverComma}; pub use path::PathStyle; diff --git a/compiler/rustc_parse/src/parser/pat.rs b/compiler/rustc_parse/src/parser/pat.rs index 9754691a0b9..c4d30b3d328 100644 --- a/compiler/rustc_parse/src/parser/pat.rs +++ b/compiler/rustc_parse/src/parser/pat.rs @@ -1516,7 +1516,7 @@ impl<'a> Parser<'a> { || self.check_noexpect(&token::DotDotDot) || self.check_keyword(exp!(Underscore)) { - etc = PatFieldsRest::Rest; + etc = PatFieldsRest::Rest(self.token.span); let mut etc_sp = self.token.span; if first_etc_and_maybe_comma_span.is_none() { if let Some(comma_tok) = |