diff options
Diffstat (limited to 'compiler')
| -rw-r--r-- | compiler/rustc_builtin_macros/src/source_util.rs | 100 | ||||
| -rw-r--r-- | compiler/rustc_driver_impl/src/lib.rs | 12 | ||||
| -rw-r--r-- | compiler/rustc_expand/src/module.rs | 9 | ||||
| -rw-r--r-- | compiler/rustc_expand/src/proc_macro_server.rs | 11 | ||||
| -rw-r--r-- | compiler/rustc_interface/src/interface.rs | 22 | ||||
| -rw-r--r-- | compiler/rustc_interface/src/passes.rs | 17 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/lexer/mod.rs | 2 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/lib.rs | 36 | ||||
| -rw-r--r-- | compiler/rustc_parse/src/parser/tests.rs | 7 |
9 files changed, 134 insertions, 82 deletions
diff --git a/compiler/rustc_builtin_macros/src/source_util.rs b/compiler/rustc_builtin_macros/src/source_util.rs index 37bab5be542..11b868f81a9 100644 --- a/compiler/rustc_builtin_macros/src/source_util.rs +++ b/compiler/rustc_builtin_macros/src/source_util.rs @@ -1,3 +1,5 @@ +//! The implementation of built-in macros which relate to the file system. + use std::path::{Path, PathBuf}; use std::rc::Rc; use std::sync::Arc; @@ -11,9 +13,11 @@ use rustc_expand::base::{ }; use rustc_expand::module::DirOwnership; use rustc_lint_defs::BuiltinLintDiag; -use rustc_parse::parser::{ForceCollect, Parser}; +use rustc_parse::lexer::StripTokens; +use rustc_parse::parser::ForceCollect; use rustc_parse::{new_parser_from_file, unwrap_or_emit_fatal, utf8_error}; use rustc_session::lint::builtin::INCOMPLETE_INCLUDE; +use rustc_session::parse::ParseSess; use rustc_span::source_map::SourceMap; use rustc_span::{ByteSymbol, Pos, Span, Symbol}; use smallvec::SmallVec; @@ -23,11 +27,7 @@ use crate::util::{ check_zero_tts, get_single_str_from_tts, get_single_str_spanned_from_tts, parse_expr, }; -// These macros all relate to the file system; they either return -// the column/row/filename of the expression, or they include -// a given file into the current one. - -/// line!(): expands to the current line number +/// Expand `line!()` to the current line number. pub(crate) fn expand_line( cx: &mut ExtCtxt<'_>, sp: Span, @@ -42,7 +42,7 @@ pub(crate) fn expand_line( ExpandResult::Ready(MacEager::expr(cx.expr_u32(topmost, loc.line as u32))) } -/* column!(): expands to the current column number */ +/// Expand `column!()` to the current column number. pub(crate) fn expand_column( cx: &mut ExtCtxt<'_>, sp: Span, @@ -57,9 +57,7 @@ pub(crate) fn expand_column( ExpandResult::Ready(MacEager::expr(cx.expr_u32(topmost, loc.col.to_usize() as u32 + 1))) } -/// file!(): expands to the current filename */ -/// The source_file (`loc.file`) contains a bunch more information we could spit -/// out if we wanted. +/// Expand `file!()` to the current filename. pub(crate) fn expand_file( cx: &mut ExtCtxt<'_>, sp: Span, @@ -81,6 +79,7 @@ pub(crate) fn expand_file( ))) } +/// Expand `stringify!($input)`. pub(crate) fn expand_stringify( cx: &mut ExtCtxt<'_>, sp: Span, @@ -91,6 +90,7 @@ pub(crate) fn expand_stringify( ExpandResult::Ready(MacEager::expr(cx.expr_str(sp, Symbol::intern(&s)))) } +/// Expand `module_path!()` to (a textual representation of) the current module path. pub(crate) fn expand_mod( cx: &mut ExtCtxt<'_>, sp: Span, @@ -104,9 +104,9 @@ pub(crate) fn expand_mod( ExpandResult::Ready(MacEager::expr(cx.expr_str(sp, Symbol::intern(&string)))) } -/// include! : parse the given file as an expr -/// This is generally a bad idea because it's going to behave -/// unhygienically. +/// Expand `include!($input)`. +/// +/// This works in item and expression position. Notably, it doesn't work in pattern position. pub(crate) fn expand_include<'cx>( cx: &'cx mut ExtCtxt<'_>, sp: Span, @@ -116,39 +116,48 @@ pub(crate) fn expand_include<'cx>( let ExpandResult::Ready(mac) = get_single_str_from_tts(cx, sp, tts, "include!") else { return ExpandResult::Retry(()); }; - let file = match mac { - Ok(file) => file, + let path = match mac { + Ok(path) => path, Err(guar) => return ExpandResult::Ready(DummyResult::any(sp, guar)), }; // The file will be added to the code map by the parser - let file = match resolve_path(&cx.sess, file.as_str(), sp) { - Ok(f) => f, + let path = match resolve_path(&cx.sess, path.as_str(), sp) { + Ok(path) => path, Err(err) => { let guar = err.emit(); return ExpandResult::Ready(DummyResult::any(sp, guar)); } }; - let p = unwrap_or_emit_fatal(new_parser_from_file(cx.psess(), &file, Some(sp))); // If in the included file we have e.g., `mod bar;`, - // then the path of `bar.rs` should be relative to the directory of `file`. + // then the path of `bar.rs` should be relative to the directory of `path`. // See https://github.com/rust-lang/rust/pull/69838/files#r395217057 for a discussion. // `MacroExpander::fully_expand_fragment` later restores, so "stack discipline" is maintained. - let dir_path = file.parent().unwrap_or(&file).to_owned(); + let dir_path = path.parent().unwrap_or(&path).to_owned(); cx.current_expansion.module = Rc::new(cx.current_expansion.module.with_dir_path(dir_path)); cx.current_expansion.dir_ownership = DirOwnership::Owned { relative: None }; struct ExpandInclude<'a> { - p: Parser<'a>, + psess: &'a ParseSess, + path: PathBuf, node_id: ast::NodeId, + span: Span, } impl<'a> MacResult for ExpandInclude<'a> { - fn make_expr(mut self: Box<ExpandInclude<'a>>) -> Option<Box<ast::Expr>> { - let expr = parse_expr(&mut self.p).ok()?; - if self.p.token != token::Eof { - self.p.psess.buffer_lint( + fn make_expr(self: Box<ExpandInclude<'a>>) -> Option<Box<ast::Expr>> { + let mut p = unwrap_or_emit_fatal(new_parser_from_file( + self.psess, + &self.path, + // Don't strip frontmatter for backward compatibility, `---` may be the start of a + // manifold negation. FIXME: Ideally, we wouldn't strip shebangs here either. + StripTokens::Shebang, + Some(self.span), + )); + let expr = parse_expr(&mut p).ok()?; + if p.token != token::Eof { + p.psess.buffer_lint( INCOMPLETE_INCLUDE, - self.p.token.span, + p.token.span, self.node_id, BuiltinLintDiag::IncompleteInclude, ); @@ -156,24 +165,27 @@ pub(crate) fn expand_include<'cx>( Some(expr) } - fn make_items(mut self: Box<ExpandInclude<'a>>) -> Option<SmallVec<[Box<ast::Item>; 1]>> { + fn make_items(self: Box<ExpandInclude<'a>>) -> Option<SmallVec<[Box<ast::Item>; 1]>> { + let mut p = unwrap_or_emit_fatal(new_parser_from_file( + self.psess, + &self.path, + StripTokens::ShebangAndFrontmatter, + Some(self.span), + )); let mut ret = SmallVec::new(); loop { - match self.p.parse_item(ForceCollect::No) { + match p.parse_item(ForceCollect::No) { Err(err) => { err.emit(); break; } Ok(Some(item)) => ret.push(item), Ok(None) => { - if self.p.token != token::Eof { - self.p - .dcx() - .create_err(errors::ExpectedItem { - span: self.p.token.span, - token: &pprust::token_to_string(&self.p.token), - }) - .emit(); + if p.token != token::Eof { + p.dcx().emit_err(errors::ExpectedItem { + span: p.token.span, + token: &pprust::token_to_string(&p.token), + }); } break; @@ -184,10 +196,17 @@ pub(crate) fn expand_include<'cx>( } } - ExpandResult::Ready(Box::new(ExpandInclude { p, node_id: cx.current_expansion.lint_node_id })) + ExpandResult::Ready(Box::new(ExpandInclude { + psess: cx.psess(), + path, + node_id: cx.current_expansion.lint_node_id, + span: sp, + })) } -/// `include_str!`: read the given file, insert it as a literal string expr +/// Expand `include_str!($input)` to the content of the UTF-8-encoded file given by path `$input` as a string literal. +/// +/// This works in expression, pattern and statement position. pub(crate) fn expand_include_str( cx: &mut ExtCtxt<'_>, sp: Span, @@ -206,6 +225,7 @@ pub(crate) fn expand_include_str( Ok((bytes, bsp)) => match std::str::from_utf8(&bytes) { Ok(src) => { let interned_src = Symbol::intern(src); + // MacEager converts the expr into a pat if need be. MacEager::expr(cx.expr_str(cx.with_def_site_ctxt(bsp), interned_src)) } Err(utf8err) => { @@ -218,6 +238,9 @@ pub(crate) fn expand_include_str( }) } +/// Expand `include_bytes!($input)` to the content of the file given by path `$input`. +/// +/// This works in expression, pattern and statement position. pub(crate) fn expand_include_bytes( cx: &mut ExtCtxt<'_>, sp: Span, @@ -237,6 +260,7 @@ pub(crate) fn expand_include_bytes( // Don't care about getting the span for the raw bytes, // because the console can't really show them anyway. let expr = cx.expr(sp, ast::ExprKind::IncludedBytes(ByteSymbol::intern(&bytes))); + // MacEager converts the expr into a pat if need be. MacEager::expr(expr) } Err(dummy) => dummy, diff --git a/compiler/rustc_driver_impl/src/lib.rs b/compiler/rustc_driver_impl/src/lib.rs index f3ed6042105..d00a4c35834 100644 --- a/compiler/rustc_driver_impl/src/lib.rs +++ b/compiler/rustc_driver_impl/src/lib.rs @@ -51,6 +51,7 @@ use rustc_lint::unerased_lint_store; use rustc_metadata::creader::MetadataLoader; use rustc_metadata::locator; use rustc_middle::ty::TyCtxt; +use rustc_parse::lexer::StripTokens; use rustc_parse::{new_parser_from_file, new_parser_from_source_str, unwrap_or_emit_fatal}; use rustc_session::config::{ CG_OPTIONS, CrateType, ErrorOutputType, Input, OptionDesc, OutFileName, OutputType, Sysroot, @@ -1288,10 +1289,15 @@ fn warn_on_confusing_output_filename_flag( fn parse_crate_attrs<'a>(sess: &'a Session) -> PResult<'a, ast::AttrVec> { let mut parser = unwrap_or_emit_fatal(match &sess.io.input { - Input::File(file) => new_parser_from_file(&sess.psess, file, None), - Input::Str { name, input } => { - new_parser_from_source_str(&sess.psess, name.clone(), input.clone()) + Input::File(file) => { + new_parser_from_file(&sess.psess, file, StripTokens::ShebangAndFrontmatter, None) } + Input::Str { name, input } => new_parser_from_source_str( + &sess.psess, + name.clone(), + input.clone(), + StripTokens::ShebangAndFrontmatter, + ), }); parser.parse_inner_attributes() } diff --git a/compiler/rustc_expand/src/module.rs b/compiler/rustc_expand/src/module.rs index 19f3cdbc549..79ab3cab22c 100644 --- a/compiler/rustc_expand/src/module.rs +++ b/compiler/rustc_expand/src/module.rs @@ -4,6 +4,7 @@ use std::path::{self, Path, PathBuf}; use rustc_ast::{AttrVec, Attribute, Inline, Item, ModSpans}; use rustc_attr_parsing::validate_attr; use rustc_errors::{Diag, ErrorGuaranteed}; +use rustc_parse::lexer::StripTokens; use rustc_parse::{exp, new_parser_from_file, unwrap_or_emit_fatal}; use rustc_session::Session; use rustc_session::parse::ParseSess; @@ -67,8 +68,12 @@ pub(crate) fn parse_external_mod( } // Actually parse the external file as a module. - let mut parser = - unwrap_or_emit_fatal(new_parser_from_file(&sess.psess, &mp.file_path, Some(span))); + let mut parser = unwrap_or_emit_fatal(new_parser_from_file( + &sess.psess, + &mp.file_path, + StripTokens::ShebangAndFrontmatter, + Some(span), + )); let (inner_attrs, items, inner_span) = parser.parse_mod(exp!(Eof)).map_err(|err| ModError::ParserError(err))?; attrs.extend(inner_attrs); diff --git a/compiler/rustc_expand/src/proc_macro_server.rs b/compiler/rustc_expand/src/proc_macro_server.rs index 5b1d3d6d35b..295573f4492 100644 --- a/compiler/rustc_expand/src/proc_macro_server.rs +++ b/compiler/rustc_expand/src/proc_macro_server.rs @@ -8,7 +8,7 @@ use rustc_ast::util::literal::escape_byte_str_symbol; use rustc_ast_pretty::pprust; use rustc_data_structures::fx::FxHashMap; use rustc_errors::{Diag, ErrorGuaranteed, MultiSpan, PResult}; -use rustc_parse::lexer::nfc_normalize; +use rustc_parse::lexer::{StripTokens, nfc_normalize}; use rustc_parse::parser::Parser; use rustc_parse::{exp, new_parser_from_source_str, source_str_to_stream, unwrap_or_emit_fatal}; use rustc_proc_macro::bridge::{ @@ -485,8 +485,13 @@ impl server::FreeFunctions for Rustc<'_, '_> { fn literal_from_str(&mut self, s: &str) -> Result<Literal<Self::Span, Self::Symbol>, ()> { let name = FileName::proc_macro_source_code(s); - let mut parser = - unwrap_or_emit_fatal(new_parser_from_source_str(self.psess(), name, s.to_owned())); + + let mut parser = unwrap_or_emit_fatal(new_parser_from_source_str( + self.psess(), + name, + s.to_owned(), + StripTokens::Nothing, + )); let first_span = parser.token.span.data(); let minus_present = parser.eat(exp!(Minus)); diff --git a/compiler/rustc_interface/src/interface.rs b/compiler/rustc_interface/src/interface.rs index 4c820b8877b..b52c5b4cd66 100644 --- a/compiler/rustc_interface/src/interface.rs +++ b/compiler/rustc_interface/src/interface.rs @@ -13,7 +13,8 @@ use rustc_lint::LintStore; use rustc_middle::ty; use rustc_middle::ty::CurrentGcx; use rustc_middle::util::Providers; -use rustc_parse::new_parser_from_simple_source_str; +use rustc_parse::lexer::StripTokens; +use rustc_parse::new_parser_from_source_str; use rustc_parse::parser::attr::AllowLeadingUnsafe; use rustc_query_impl::QueryCtxt; use rustc_query_system::query::print_query_stack; @@ -68,7 +69,8 @@ pub(crate) fn parse_cfg(dcx: DiagCtxtHandle<'_>, cfgs: Vec<String>) -> Cfg { }; } - match new_parser_from_simple_source_str(&psess, filename, s.to_string()) { + match new_parser_from_source_str(&psess, filename, s.to_string(), StripTokens::Nothing) + { Ok(mut parser) => match parser.parse_meta_item(AllowLeadingUnsafe::No) { Ok(meta_item) if parser.token == token::Eof => { if meta_item.path.segments.len() != 1 { @@ -166,13 +168,15 @@ pub(crate) fn parse_check_cfg(dcx: DiagCtxtHandle<'_>, specs: Vec<String>) -> Ch error!("expected `cfg(name, values(\"value1\", \"value2\", ... \"valueN\"))`") }; - let mut parser = match new_parser_from_simple_source_str(&psess, filename, s.to_string()) { - Ok(parser) => parser, - Err(errs) => { - errs.into_iter().for_each(|err| err.cancel()); - expected_error(); - } - }; + let mut parser = + match new_parser_from_source_str(&psess, filename, s.to_string(), StripTokens::Nothing) + { + Ok(parser) => parser, + Err(errs) => { + errs.into_iter().for_each(|err| err.cancel()); + expected_error(); + } + }; let meta_item = match parser.parse_meta_item(AllowLeadingUnsafe::No) { Ok(meta_item) if parser.token == token::Eof => meta_item, diff --git a/compiler/rustc_interface/src/passes.rs b/compiler/rustc_interface/src/passes.rs index 6d9751d7d4d..d39219bfd66 100644 --- a/compiler/rustc_interface/src/passes.rs +++ b/compiler/rustc_interface/src/passes.rs @@ -28,6 +28,7 @@ use rustc_middle::arena::Arena; use rustc_middle::dep_graph::DepsType; use rustc_middle::ty::{self, CurrentGcx, GlobalCtxt, RegisteredTools, TyCtxt}; use rustc_middle::util::Providers; +use rustc_parse::lexer::StripTokens; use rustc_parse::{new_parser_from_file, new_parser_from_source_str, unwrap_or_emit_fatal}; use rustc_passes::{abi_test, input_stats, layout_test}; use rustc_resolve::{Resolver, ResolverOutputs}; @@ -52,10 +53,18 @@ pub fn parse<'a>(sess: &'a Session) -> ast::Crate { let mut krate = sess .time("parse_crate", || { let mut parser = unwrap_or_emit_fatal(match &sess.io.input { - Input::File(file) => new_parser_from_file(&sess.psess, file, None), - Input::Str { input, name } => { - new_parser_from_source_str(&sess.psess, name.clone(), input.clone()) - } + Input::File(file) => new_parser_from_file( + &sess.psess, + file, + StripTokens::ShebangAndFrontmatter, + None, + ), + Input::Str { input, name } => new_parser_from_source_str( + &sess.psess, + name.clone(), + input.clone(), + StripTokens::ShebangAndFrontmatter, + ), }); parser.parse_crate_mod() }) diff --git a/compiler/rustc_parse/src/lexer/mod.rs b/compiler/rustc_parse/src/lexer/mod.rs index f5f081efc49..51019db7c00 100644 --- a/compiler/rustc_parse/src/lexer/mod.rs +++ b/compiler/rustc_parse/src/lexer/mod.rs @@ -45,7 +45,7 @@ pub(crate) struct UnmatchedDelim { } /// Which tokens should be stripped before lexing the tokens. -pub(crate) enum StripTokens { +pub enum StripTokens { /// Strip both shebang and frontmatter. ShebangAndFrontmatter, /// Strip the shebang but not frontmatter. diff --git a/compiler/rustc_parse/src/lib.rs b/compiler/rustc_parse/src/lib.rs index d8792d7af4c..88b67d792de 100644 --- a/compiler/rustc_parse/src/lib.rs +++ b/compiler/rustc_parse/src/lib.rs @@ -54,29 +54,18 @@ pub fn unwrap_or_emit_fatal<T>(expr: Result<T, Vec<Diag<'_>>>) -> T { } } -/// Creates a new parser from a 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_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::ShebangAndFrontmatter) -} - -/// Creates a new parser from a simple (no shebang, no frontmatter) source string. +/// Creates a new parser from a 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( +pub fn new_parser_from_source_str( psess: &ParseSess, name: FileName, source: String, + strip_tokens: StripTokens, ) -> 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) + new_parser_from_source_file(psess, source_file, strip_tokens) } /// Creates a new parser from a filename. On failure, the errors must be consumed via @@ -87,6 +76,7 @@ pub fn new_parser_from_simple_source_str( pub fn new_parser_from_file<'a>( psess: &'a ParseSess, path: &Path, + strip_tokens: StripTokens, sp: Option<Span>, ) -> Result<Parser<'a>, Vec<Diag<'a>>> { let sm = psess.source_map(); @@ -110,7 +100,7 @@ pub fn new_parser_from_file<'a>( } err.emit(); }); - new_parser_from_source_file(psess, source_file, StripTokens::ShebangAndFrontmatter) + new_parser_from_source_file(psess, source_file, strip_tokens) } pub fn utf8_error<E: EmissionGuarantee>( @@ -172,6 +162,9 @@ fn new_parser_from_source_file( Ok(parser) } +/// Given a source string, produces a sequence of token trees. +/// +/// NOTE: This only strips shebangs, not frontmatter! pub fn source_str_to_stream( psess: &ParseSess, name: FileName, @@ -179,13 +172,16 @@ pub fn source_str_to_stream( override_span: Option<Span>, ) -> 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, but for compatibility reason still strip the shebang + // FIXME(frontmatter): Consider stripping frontmatter in a future edition. We can't strip them + // in the current edition since that would be breaking. + // See also <https://github.com/rust-lang/rust/issues/145520>. + // Alternatively, stop stripping shebangs here, too, if T-lang and crater approve. 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 -/// parsing the token stream. +/// Given a source file, produces a sequence of token trees. +/// +/// Returns any buffered errors from parsing the token stream. fn source_file_to_stream<'psess>( psess: &'psess ParseSess, source_file: Arc<SourceFile>, diff --git a/compiler/rustc_parse/src/parser/tests.rs b/compiler/rustc_parse/src/parser/tests.rs index a6e7266e71b..e645fb47b9e 100644 --- a/compiler/rustc_parse/src/parser/tests.rs +++ b/compiler/rustc_parse/src/parser/tests.rs @@ -22,6 +22,7 @@ use rustc_span::{ }; use termcolor::WriteColor; +use crate::lexer::StripTokens; use crate::parser::{ForceCollect, Parser}; use crate::{new_parser_from_source_str, source_str_to_stream, unwrap_or_emit_fatal}; @@ -35,6 +36,7 @@ fn string_to_parser(psess: &ParseSess, source_str: String) -> Parser<'_> { psess, PathBuf::from("bogofile").into(), source_str, + StripTokens::Nothing, )) } @@ -2240,7 +2242,7 @@ fn parse_item_from_source_str( source: String, psess: &ParseSess, ) -> PResult<'_, Option<Box<ast::Item>>> { - unwrap_or_emit_fatal(new_parser_from_source_str(psess, name, source)) + unwrap_or_emit_fatal(new_parser_from_source_str(psess, name, source, StripTokens::Nothing)) .parse_item(ForceCollect::No) } @@ -2520,7 +2522,8 @@ fn ttdelim_span() { source: String, psess: &ParseSess, ) -> PResult<'_, Box<ast::Expr>> { - unwrap_or_emit_fatal(new_parser_from_source_str(psess, name, source)).parse_expr() + unwrap_or_emit_fatal(new_parser_from_source_str(psess, name, source, StripTokens::Nothing)) + .parse_expr() } create_default_session_globals_then(|| { |