diff options
| author | bors <bors@rust-lang.org> | 2025-05-06 06:37:30 +0000 |
|---|---|---|
| committer | bors <bors@rust-lang.org> | 2025-05-06 06:37:30 +0000 |
| commit | 651e9cf327358b28db7e37a2ae61727f4a2ef232 (patch) | |
| tree | 9e08a2741b303e04c0f4b98a180efd01d8716f1d /compiler | |
| parent | 7295b08a17d1107155acd4b552069e3705b0ab1f (diff) | |
| parent | 546c1c2dd48ba6eded56a9ee74d78cab8e7ad204 (diff) | |
| download | rust-651e9cf327358b28db7e37a2ae61727f4a2ef232.tar.gz rust-651e9cf327358b28db7e37a2ae61727f4a2ef232.zip | |
Auto merge of #140695 - Zalathar:rollup-i32gzbo, r=Zalathar
Rollup of 12 pull requests Successful merges: - #139550 (Fix `-Zremap-path-scope` rmeta handling) - #139764 (Consistent trait bounds for ExtractIf Debug impls) - #139773 (Implement `Iterator::last` for `vec::IntoIter`) - #140035 (Implement RFC 3503: frontmatters) - #140251 (coverage-dump: Resolve global file IDs to filenames) - #140393 (std: get rid of `sys_common::process`) - #140532 (Fix RustAnalyzer discovery of rustc's `stable_mir` crate) - #140598 (Steer docs to `utf8_chunks` and `Iterator::take`) - #140634 (Use more accurate ELF flags on MIPS) - #140673 (Clean rustdoc tests folder) - #140678 (Be a bit more relaxed about not yet constrained infer vars in closure upvar analysis) - #140687 (Update mdbook to 0.4.49) r? `@ghost` `@rustbot` modify labels: rollup
Diffstat (limited to 'compiler')
28 files changed, 565 insertions, 137 deletions
diff --git a/compiler/rustc_ast_passes/src/feature_gate.rs b/compiler/rustc_ast_passes/src/feature_gate.rs index e312f15f05b..915613a3913 100644 --- a/compiler/rustc_ast_passes/src/feature_gate.rs +++ b/compiler/rustc_ast_passes/src/feature_gate.rs @@ -514,6 +514,7 @@ pub fn check_crate(krate: &ast::Crate, sess: &Session, features: &Features) { gate_all!(contracts_internals, "contract internal machinery is for internal use only"); gate_all!(where_clause_attrs, "attributes in `where` clause are unstable"); gate_all!(super_let, "`super let` is experimental"); + gate_all!(frontmatter, "frontmatters are experimental"); if !visitor.features.never_patterns() { if let Some(spans) = spans.get(&sym::never_patterns) { diff --git a/compiler/rustc_codegen_ssa/src/back/metadata.rs b/compiler/rustc_codegen_ssa/src/back/metadata.rs index ebcccf1b97d..ec46c71b0e4 100644 --- a/compiler/rustc_codegen_ssa/src/back/metadata.rs +++ b/compiler/rustc_codegen_ssa/src/back/metadata.rs @@ -270,45 +270,61 @@ pub(super) fn elf_os_abi(sess: &Session) -> u8 { pub(super) fn elf_e_flags(architecture: Architecture, sess: &Session) -> u32 { match architecture { - Architecture::Mips => { - let arch = match sess.target.options.cpu.as_ref() { - "mips1" => elf::EF_MIPS_ARCH_1, - "mips2" => elf::EF_MIPS_ARCH_2, + Architecture::Mips | Architecture::Mips64 | Architecture::Mips64_N32 => { + // "N32" indicates an "ILP32" data model on a 64-bit MIPS CPU + // like SPARC's "v8+", x86_64's "x32", or the watchOS "arm64_32". + let is_32bit = architecture == Architecture::Mips; + let mut e_flags = match sess.target.options.cpu.as_ref() { + "mips1" if is_32bit => elf::EF_MIPS_ARCH_1, + "mips2" if is_32bit => elf::EF_MIPS_ARCH_2, "mips3" => elf::EF_MIPS_ARCH_3, "mips4" => elf::EF_MIPS_ARCH_4, "mips5" => elf::EF_MIPS_ARCH_5, - s if s.contains("r6") => elf::EF_MIPS_ARCH_32R6, - _ => elf::EF_MIPS_ARCH_32R2, + "mips32r2" if is_32bit => elf::EF_MIPS_ARCH_32R2, + "mips32r6" if is_32bit => elf::EF_MIPS_ARCH_32R6, + "mips64r2" if !is_32bit => elf::EF_MIPS_ARCH_64R2, + "mips64r6" if !is_32bit => elf::EF_MIPS_ARCH_64R6, + s if s.starts_with("mips32") && !is_32bit => { + sess.dcx().fatal(format!("invalid CPU `{}` for 64-bit MIPS target", s)) + } + s if s.starts_with("mips64") && is_32bit => { + sess.dcx().fatal(format!("invalid CPU `{}` for 32-bit MIPS target", s)) + } + _ if is_32bit => elf::EF_MIPS_ARCH_32R2, + _ => elf::EF_MIPS_ARCH_64R2, }; - let mut e_flags = elf::EF_MIPS_CPIC | arch; - - // If the ABI is explicitly given, use it or default to O32. - match sess.target.options.llvm_abiname.to_lowercase().as_str() { - "n32" => e_flags |= elf::EF_MIPS_ABI2, - "o32" => e_flags |= elf::EF_MIPS_ABI_O32, - _ => e_flags |= elf::EF_MIPS_ABI_O32, + // If the ABI is explicitly given, use it, or default to O32 on 32-bit MIPS, + // which is the only "true" 32-bit option that LLVM supports. + match sess.target.options.llvm_abiname.as_ref() { + "o32" if is_32bit => e_flags |= elf::EF_MIPS_ABI_O32, + "n32" if !is_32bit => e_flags |= elf::EF_MIPS_ABI2, + "n64" if !is_32bit => {} + "" if is_32bit => e_flags |= elf::EF_MIPS_ABI_O32, + "" => sess.dcx().fatal("LLVM ABI must be specifed for 64-bit MIPS targets"), + s if is_32bit => { + sess.dcx().fatal(format!("invalid LLVM ABI `{}` for 32-bit MIPS target", s)) + } + s => sess.dcx().fatal(format!("invalid LLVM ABI `{}` for 64-bit MIPS target", s)), }; if sess.target.options.relocation_model != RelocModel::Static { - e_flags |= elf::EF_MIPS_PIC; + // PIC means position-independent code. CPIC means "calls PIC". + // CPIC was mutually exclusive with PIC according to + // the SVR4 MIPS ABI https://refspecs.linuxfoundation.org/elf/mipsabi.pdf + // and should have only appeared on static objects with dynamically calls. + // At some point someone (GCC?) decided to set CPIC even for PIC. + // Nowadays various things expect both set on the same object file + // and may even error if you mix CPIC and non-CPIC object files, + // despite that being the entire point of the CPIC ABI extension! + // As we are in Rome, we do as the Romans do. + e_flags |= elf::EF_MIPS_PIC | elf::EF_MIPS_CPIC; } if sess.target.options.cpu.contains("r6") { e_flags |= elf::EF_MIPS_NAN2008; } e_flags } - Architecture::Mips64 => { - // copied from `mips64el-linux-gnuabi64-gcc foo.c -c` - let e_flags = elf::EF_MIPS_CPIC - | elf::EF_MIPS_PIC - | if sess.target.options.cpu.contains("r6") { - elf::EF_MIPS_ARCH_64R6 | elf::EF_MIPS_NAN2008 - } else { - elf::EF_MIPS_ARCH_64R2 - }; - e_flags - } Architecture::Riscv32 | Architecture::Riscv64 => { // Source: https://github.com/riscv-non-isa/riscv-elf-psabi-doc/blob/079772828bd10933d34121117a222b4cc0ee2200/riscv-elf.adoc let mut e_flags: u32 = 0x0; diff --git a/compiler/rustc_driver_impl/Cargo.toml b/compiler/rustc_driver_impl/Cargo.toml index c823d11126e..9da4f2dbc27 100644 --- a/compiler/rustc_driver_impl/Cargo.toml +++ b/compiler/rustc_driver_impl/Cargo.toml @@ -44,13 +44,13 @@ rustc_privacy = { path = "../rustc_privacy" } rustc_query_system = { path = "../rustc_query_system" } rustc_resolve = { path = "../rustc_resolve" } rustc_session = { path = "../rustc_session" } -rustc_smir = { path = "../rustc_smir" } rustc_span = { path = "../rustc_span" } rustc_target = { path = "../rustc_target" } rustc_trait_selection = { path = "../rustc_trait_selection" } rustc_ty_utils = { path = "../rustc_ty_utils" } serde_json = "1.0.59" shlex = "1.0" +stable_mir = { path = "../stable_mir", features = ["rustc_internal"] } tracing = { version = "0.1.35" } # tidy-alphabetical-end diff --git a/compiler/rustc_driver_impl/src/pretty.rs b/compiler/rustc_driver_impl/src/pretty.rs index 16d70af7e05..ec77043cd12 100644 --- a/compiler/rustc_driver_impl/src/pretty.rs +++ b/compiler/rustc_driver_impl/src/pretty.rs @@ -10,8 +10,8 @@ use rustc_middle::ty::{self, TyCtxt}; use rustc_mir_build::thir::print::{thir_flat, thir_tree}; use rustc_session::Session; use rustc_session::config::{OutFileName, PpHirMode, PpMode, PpSourceMode}; -use rustc_smir::rustc_internal::pretty::write_smir_pretty; use rustc_span::{FileName, Ident}; +use stable_mir::rustc_internal::pretty::write_smir_pretty; use tracing::debug; use {rustc_ast as ast, rustc_hir_pretty as pprust_hir}; diff --git a/compiler/rustc_feature/src/unstable.rs b/compiler/rustc_feature/src/unstable.rs index f1bc2c5ea88..f3412159a8a 100644 --- a/compiler/rustc_feature/src/unstable.rs +++ b/compiler/rustc_feature/src/unstable.rs @@ -510,6 +510,8 @@ declare_features! ( (incomplete, fn_delegation, "1.76.0", Some(118212)), /// Allows impls for the Freeze trait. (internal, freeze_impls, "1.78.0", Some(121675)), + /// Frontmatter `---` blocks for use by external tools. + (unstable, frontmatter, "CURRENT_RUSTC_VERSION", Some(136889)), /// Allows defining gen blocks and `gen fn`. (unstable, gen_blocks, "1.75.0", Some(117078)), /// Infer generic args for both consts and types. diff --git a/compiler/rustc_hir_typeck/src/expr_use_visitor.rs b/compiler/rustc_hir_typeck/src/expr_use_visitor.rs index 17e13ec0a37..3493d359028 100644 --- a/compiler/rustc_hir_typeck/src/expr_use_visitor.rs +++ b/compiler/rustc_hir_typeck/src/expr_use_visitor.rs @@ -158,7 +158,7 @@ pub trait TypeInformationCtxt<'tcx> { fn resolve_vars_if_possible<T: TypeFoldable<TyCtxt<'tcx>>>(&self, t: T) -> T; - fn try_structurally_resolve_type(&self, span: Span, ty: Ty<'tcx>) -> Ty<'tcx>; + fn structurally_resolve_type(&self, span: Span, ty: Ty<'tcx>) -> Ty<'tcx>; fn report_bug(&self, span: Span, msg: impl ToString) -> Self::Error; @@ -191,8 +191,8 @@ impl<'tcx> TypeInformationCtxt<'tcx> for &FnCtxt<'_, 'tcx> { self.infcx.resolve_vars_if_possible(t) } - fn try_structurally_resolve_type(&self, sp: Span, ty: Ty<'tcx>) -> Ty<'tcx> { - (**self).try_structurally_resolve_type(sp, ty) + fn structurally_resolve_type(&self, sp: Span, ty: Ty<'tcx>) -> Ty<'tcx> { + (**self).structurally_resolve_type(sp, ty) } fn report_bug(&self, span: Span, msg: impl ToString) -> Self::Error { @@ -236,7 +236,7 @@ impl<'tcx> TypeInformationCtxt<'tcx> for (&LateContext<'tcx>, LocalDefId) { self.0.maybe_typeck_results().expect("expected typeck results") } - fn try_structurally_resolve_type(&self, _span: Span, ty: Ty<'tcx>) -> Ty<'tcx> { + fn structurally_resolve_type(&self, _span: Span, ty: Ty<'tcx>) -> Ty<'tcx> { // FIXME: Maybe need to normalize here. ty } @@ -776,7 +776,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx // Select just those fields of the `with` // expression that will actually be used - match self.cx.try_structurally_resolve_type(with_expr.span, with_place.place.ty()).kind() { + match self.cx.structurally_resolve_type(with_expr.span, with_place.place.ty()).kind() { ty::Adt(adt, args) if adt.is_struct() => { // Consume those fields of the with expression that are needed. for (f_index, with_field) in adt.non_enum_variant().fields.iter_enumerated() { @@ -1176,7 +1176,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx /// two operations: a dereference to reach the array data and then an index to /// jump forward to the relevant item. impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx, Cx, D> { - fn resolve_type_vars_or_bug( + fn expect_and_resolve_type( &self, id: HirId, ty: Option<Ty<'tcx>>, @@ -1185,12 +1185,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx Some(ty) => { let ty = self.cx.resolve_vars_if_possible(ty); self.cx.error_reported_in_ty(ty)?; - if ty.is_ty_var() { - debug!("resolve_type_vars_or_bug: infer var from {:?}", ty); - Err(self.cx.report_bug(self.cx.tcx().hir_span(id), "encountered type variable")) - } else { - Ok(ty) - } + Ok(ty) } None => { // FIXME: We shouldn't be relying on the infcx being tainted. @@ -1201,15 +1196,15 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx } fn node_ty(&self, hir_id: HirId) -> Result<Ty<'tcx>, Cx::Error> { - self.resolve_type_vars_or_bug(hir_id, self.cx.typeck_results().node_type_opt(hir_id)) + self.expect_and_resolve_type(hir_id, self.cx.typeck_results().node_type_opt(hir_id)) } fn expr_ty(&self, expr: &hir::Expr<'_>) -> Result<Ty<'tcx>, Cx::Error> { - self.resolve_type_vars_or_bug(expr.hir_id, self.cx.typeck_results().expr_ty_opt(expr)) + self.expect_and_resolve_type(expr.hir_id, self.cx.typeck_results().expr_ty_opt(expr)) } fn expr_ty_adjusted(&self, expr: &hir::Expr<'_>) -> Result<Ty<'tcx>, Cx::Error> { - self.resolve_type_vars_or_bug( + self.expect_and_resolve_type( expr.hir_id, self.cx.typeck_results().expr_ty_adjusted_opt(expr), ) @@ -1264,10 +1259,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx // a bind-by-ref means that the base_ty will be the type of the ident itself, // but what we want here is the type of the underlying value being borrowed. // So peel off one-level, turning the &T into T. - match self - .cx - .try_structurally_resolve_type(pat.span, base_ty) - .builtin_deref(false) + match self.cx.structurally_resolve_type(pat.span, base_ty).builtin_deref(false) { Some(ty) => Ok(ty), None => { @@ -1513,10 +1505,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx if node_ty != place_ty && self .cx - .try_structurally_resolve_type( - self.cx.tcx().hir_span(base_place.hir_id), - place_ty, - ) + .structurally_resolve_type(self.cx.tcx().hir_span(base_place.hir_id), place_ty) .is_impl_trait() { projections.push(Projection { kind: ProjectionKind::OpaqueCast, ty: node_ty }); @@ -1538,7 +1527,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx let base_ty = self.expr_ty_adjusted(base)?; let ty::Ref(region, _, mutbl) = - *self.cx.try_structurally_resolve_type(base.span, base_ty).kind() + *self.cx.structurally_resolve_type(base.span, base_ty).kind() else { span_bug!(expr.span, "cat_overloaded_place: base is not a reference"); }; @@ -1556,7 +1545,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx let base_curr_ty = base_place.place.ty(); let deref_ty = match self .cx - .try_structurally_resolve_type(self.cx.tcx().hir_span(base_place.hir_id), base_curr_ty) + .structurally_resolve_type(self.cx.tcx().hir_span(base_place.hir_id), base_curr_ty) .builtin_deref(true) { Some(ty) => ty, @@ -1584,7 +1573,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx ) -> Result<VariantIdx, Cx::Error> { let res = self.cx.typeck_results().qpath_res(qpath, pat_hir_id); let ty = self.cx.typeck_results().node_type(pat_hir_id); - let ty::Adt(adt_def, _) = self.cx.try_structurally_resolve_type(span, ty).kind() else { + let ty::Adt(adt_def, _) = self.cx.structurally_resolve_type(span, ty).kind() else { return Err(self .cx .report_bug(span, "struct or tuple struct pattern not applied to an ADT")); @@ -1616,7 +1605,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx span: Span, ) -> Result<usize, Cx::Error> { let ty = self.cx.typeck_results().node_type(pat_hir_id); - match self.cx.try_structurally_resolve_type(span, ty).kind() { + match self.cx.structurally_resolve_type(span, ty).kind() { ty::Adt(adt_def, _) => Ok(adt_def.variant(variant_index).fields.len()), _ => { self.cx @@ -1631,7 +1620,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx /// Here `pat_hir_id` is the HirId of the pattern itself. fn total_fields_in_tuple(&self, pat_hir_id: HirId, span: Span) -> Result<usize, Cx::Error> { let ty = self.cx.typeck_results().node_type(pat_hir_id); - match self.cx.try_structurally_resolve_type(span, ty).kind() { + match self.cx.structurally_resolve_type(span, ty).kind() { ty::Tuple(args) => Ok(args.len()), _ => Err(self.cx.report_bug(span, "tuple pattern not applied to a tuple")), } @@ -1820,7 +1809,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx PatKind::Slice(before, ref slice, after) => { let Some(element_ty) = self .cx - .try_structurally_resolve_type(pat.span, place_with_id.place.ty()) + .structurally_resolve_type(pat.span, place_with_id.place.ty()) .builtin_index() else { debug!("explicit index of non-indexable type {:?}", place_with_id); @@ -1890,7 +1879,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx } fn is_multivariant_adt(&self, ty: Ty<'tcx>, span: Span) -> bool { - if let ty::Adt(def, _) = self.cx.try_structurally_resolve_type(span, ty).kind() { + if let ty::Adt(def, _) = self.cx.structurally_resolve_type(span, ty).kind() { // Note that if a non-exhaustive SingleVariant is defined in another crate, we need // to assume that more cases will be added to the variant in the future. This mean // that we should handle non-exhaustive SingleVariant the same way we would handle diff --git a/compiler/rustc_lexer/src/cursor.rs b/compiler/rustc_lexer/src/cursor.rs index e0e3bd0e30b..526693d3de1 100644 --- a/compiler/rustc_lexer/src/cursor.rs +++ b/compiler/rustc_lexer/src/cursor.rs @@ -1,5 +1,10 @@ use std::str::Chars; +pub enum FrontmatterAllowed { + Yes, + No, +} + /// Peekable iterator over a char sequence. /// /// Next characters can be peeked via `first` method, @@ -8,6 +13,7 @@ pub struct Cursor<'a> { len_remaining: usize, /// Iterator over chars. Slightly faster than a &str. chars: Chars<'a>, + pub(crate) frontmatter_allowed: FrontmatterAllowed, #[cfg(debug_assertions)] prev: char, } @@ -15,10 +21,11 @@ pub struct Cursor<'a> { pub(crate) const EOF_CHAR: char = '\0'; impl<'a> Cursor<'a> { - pub fn new(input: &'a str) -> Cursor<'a> { + pub fn new(input: &'a str, frontmatter_allowed: FrontmatterAllowed) -> Cursor<'a> { Cursor { len_remaining: input.len(), chars: input.chars(), + frontmatter_allowed, #[cfg(debug_assertions)] prev: EOF_CHAR, } @@ -95,6 +102,11 @@ impl<'a> Cursor<'a> { Some(c) } + /// Moves to a substring by a number of bytes. + pub(crate) fn bump_bytes(&mut self, n: usize) { + self.chars = self.as_str()[n..].chars(); + } + /// Eats symbols while predicate returns true or until the end of file is reached. pub(crate) fn eat_while(&mut self, mut predicate: impl FnMut(char) -> bool) { // It was tried making optimized version of this for eg. line comments, but diff --git a/compiler/rustc_lexer/src/lib.rs b/compiler/rustc_lexer/src/lib.rs index f9c71b2fa65..2374f388250 100644 --- a/compiler/rustc_lexer/src/lib.rs +++ b/compiler/rustc_lexer/src/lib.rs @@ -35,8 +35,8 @@ pub use unicode_xid::UNICODE_VERSION as UNICODE_XID_VERSION; use self::LiteralKind::*; use self::TokenKind::*; -pub use crate::cursor::Cursor; use crate::cursor::EOF_CHAR; +pub use crate::cursor::{Cursor, FrontmatterAllowed}; /// Parsed token. /// It doesn't contain information about data that has been parsed, @@ -57,17 +57,27 @@ impl Token { #[derive(Clone, Copy, Debug, PartialEq, Eq)] pub enum TokenKind { /// A line comment, e.g. `// comment`. - LineComment { doc_style: Option<DocStyle> }, + LineComment { + doc_style: Option<DocStyle>, + }, /// A block comment, e.g. `/* block comment */`. /// /// Block comments can be recursive, so a sequence like `/* /* */` /// will not be considered terminated and will result in a parsing error. - BlockComment { doc_style: Option<DocStyle>, terminated: bool }, + BlockComment { + doc_style: Option<DocStyle>, + terminated: bool, + }, /// Any whitespace character sequence. Whitespace, + Frontmatter { + has_invalid_preceding_whitespace: bool, + invalid_infostring: bool, + }, + /// An identifier or keyword, e.g. `ident` or `continue`. Ident, @@ -109,10 +119,15 @@ pub enum TokenKind { /// this type will need to check for and reject that case. /// /// See [LiteralKind] for more details. - Literal { kind: LiteralKind, suffix_start: u32 }, + Literal { + kind: LiteralKind, + suffix_start: u32, + }, /// A lifetime, e.g. `'a`. - Lifetime { starts_with_number: bool }, + Lifetime { + starts_with_number: bool, + }, /// `;` Semi, @@ -280,7 +295,7 @@ pub fn strip_shebang(input: &str) -> Option<usize> { #[inline] pub fn validate_raw_str(input: &str, prefix_len: u32) -> Result<(), RawStrError> { debug_assert!(!input.is_empty()); - let mut cursor = Cursor::new(input); + let mut cursor = Cursor::new(input, FrontmatterAllowed::No); // Move past the leading `r` or `br`. for _ in 0..prefix_len { cursor.bump().unwrap(); @@ -290,7 +305,7 @@ pub fn validate_raw_str(input: &str, prefix_len: u32) -> Result<(), RawStrError> /// Creates an iterator that produces tokens from the input string. pub fn tokenize(input: &str) -> impl Iterator<Item = Token> { - let mut cursor = Cursor::new(input); + let mut cursor = Cursor::new(input, FrontmatterAllowed::No); std::iter::from_fn(move || { let token = cursor.advance_token(); if token.kind != TokenKind::Eof { Some(token) } else { None } @@ -361,7 +376,34 @@ impl Cursor<'_> { Some(c) => c, None => return Token::new(TokenKind::Eof, 0), }; + let token_kind = match first_char { + c if matches!(self.frontmatter_allowed, FrontmatterAllowed::Yes) + && is_whitespace(c) => + { + let mut last = first_char; + while is_whitespace(self.first()) { + let Some(c) = self.bump() else { + break; + }; + last = c; + } + // invalid frontmatter opening as whitespace preceding it isn't newline. + // combine the whitespace and the frontmatter to a single token as we shall + // error later. + if last != '\n' && self.as_str().starts_with("---") { + self.bump(); + self.frontmatter(true) + } else { + Whitespace + } + } + '-' if matches!(self.frontmatter_allowed, FrontmatterAllowed::Yes) + && self.as_str().starts_with("--") => + { + // happy path + self.frontmatter(false) + } // Slash, comment or block comment. '/' => match self.first() { '/' => self.line_comment(), @@ -464,11 +506,110 @@ impl Cursor<'_> { c if !c.is_ascii() && c.is_emoji_char() => self.invalid_ident(), _ => Unknown, }; + if matches!(self.frontmatter_allowed, FrontmatterAllowed::Yes) + && !matches!(token_kind, Whitespace) + { + // stop allowing frontmatters after first non-whitespace token + self.frontmatter_allowed = FrontmatterAllowed::No; + } let res = Token::new(token_kind, self.pos_within_token()); self.reset_pos_within_token(); res } + /// Given that one `-` was eaten, eat the rest of the frontmatter. + fn frontmatter(&mut self, has_invalid_preceding_whitespace: bool) -> TokenKind { + debug_assert_eq!('-', self.prev()); + + let pos = self.pos_within_token(); + self.eat_while(|c| c == '-'); + + // one `-` is eaten by the caller. + let length_opening = self.pos_within_token() - pos + 1; + + // must be ensured by the caller + debug_assert!(length_opening >= 3); + + // whitespace between the opening and the infostring. + self.eat_while(|ch| ch != '\n' && is_whitespace(ch)); + + // copied from `eat_identifier`, but allows `.` in infostring to allow something like + // `---Cargo.toml` as a valid opener + if is_id_start(self.first()) { + self.bump(); + self.eat_while(|c| is_id_continue(c) || c == '.'); + } + + self.eat_while(|ch| ch != '\n' && is_whitespace(ch)); + let invalid_infostring = self.first() != '\n'; + + let mut s = self.as_str(); + let mut found = false; + while let Some(closing) = s.find(&"-".repeat(length_opening as usize)) { + let preceding_chars_start = s[..closing].rfind("\n").map_or(0, |i| i + 1); + if s[preceding_chars_start..closing].chars().all(is_whitespace) { + // candidate found + self.bump_bytes(closing); + // in case like + // ---cargo + // --- blahblah + // or + // ---cargo + // ---- + // combine those stuff into this frontmatter token such that it gets detected later. + self.eat_until(b'\n'); + found = true; + break; + } else { + s = &s[closing + length_opening as usize..]; + } + } + + if !found { + // recovery strategy: a closing statement might have precending whitespace/newline + // but not have enough dashes to properly close. In this case, we eat until there, + // and report a mismatch in the parser. + let mut rest = self.as_str(); + // We can look for a shorter closing (starting with four dashes but closing with three) + // and other indications that Rust has started and the infostring has ended. + let mut potential_closing = rest + .find("\n---") + // n.b. only in the case where there are dashes, we move the index to the line where + // the dashes start as we eat to include that line. For other cases those are Rust code + // and not included in the frontmatter. + .map(|x| x + 1) + .or_else(|| rest.find("\nuse ")) + .or_else(|| rest.find("\n//!")) + .or_else(|| rest.find("\n#![")); + + if potential_closing.is_none() { + // a less fortunate recovery if all else fails which finds any dashes preceded by whitespace + // on a standalone line. Might be wrong. + while let Some(closing) = rest.find("---") { + let preceding_chars_start = rest[..closing].rfind("\n").map_or(0, |i| i + 1); + if rest[preceding_chars_start..closing].chars().all(is_whitespace) { + // candidate found + potential_closing = Some(closing); + break; + } else { + rest = &rest[closing + 3..]; + } + } + } + + if let Some(potential_closing) = potential_closing { + // bump to the potential closing, and eat everything on that line. + self.bump_bytes(potential_closing); + self.eat_until(b'\n'); + } else { + // eat everything. this will get reported as an unclosed frontmatter. + self.eat_while(|_| true); + } + } + + Frontmatter { has_invalid_preceding_whitespace, invalid_infostring } + } + fn line_comment(&mut self) -> TokenKind { debug_assert!(self.prev() == '/' && self.first() == '/'); self.bump(); diff --git a/compiler/rustc_lexer/src/tests.rs b/compiler/rustc_lexer/src/tests.rs index 8203ae70b07..fc8d9b9d57b 100644 --- a/compiler/rustc_lexer/src/tests.rs +++ b/compiler/rustc_lexer/src/tests.rs @@ -4,7 +4,7 @@ use super::*; fn check_raw_str(s: &str, expected: Result<u8, RawStrError>) { let s = &format!("r{}", s); - let mut cursor = Cursor::new(s); + let mut cursor = Cursor::new(s, FrontmatterAllowed::No); cursor.bump(); let res = cursor.raw_double_quoted_string(0); assert_eq!(res, expected); diff --git a/compiler/rustc_metadata/src/rmeta/encoder.rs b/compiler/rustc_metadata/src/rmeta/encoder.rs index bbff570d6c6..7ac72ef814a 100644 --- a/compiler/rustc_metadata/src/rmeta/encoder.rs +++ b/compiler/rustc_metadata/src/rmeta/encoder.rs @@ -551,8 +551,6 @@ impl<'a, 'tcx> EncodeContext<'a, 'tcx> { match source_file.name { FileName::Real(ref original_file_name) => { - // FIXME: This should probably to conditionally remapped under - // a RemapPathScopeComponents but which one? let adapted_file_name = source_map .path_mapping() .to_embeddable_absolute_path(original_file_name.clone(), working_directory); diff --git a/compiler/rustc_parse/messages.ftl b/compiler/rustc_parse/messages.ftl index ac4f7ed64e2..3e953e6c855 100644 --- a/compiler/rustc_parse/messages.ftl +++ b/compiler/rustc_parse/messages.ftl @@ -297,6 +297,19 @@ parse_forgot_paren = perhaps you forgot parentheses? parse_found_expr_would_be_stmt = expected expression, found `{$token}` .label = expected expression +parse_frontmatter_extra_characters_after_close = extra characters after frontmatter close are not allowed +parse_frontmatter_invalid_close_preceding_whitespace = invalid preceding whitespace for frontmatter close + .note = frontmatter close should not be preceded by whitespace +parse_frontmatter_invalid_infostring = invalid infostring for frontmatter + .note = frontmatter infostrings must be a single identifier immediately following the opening +parse_frontmatter_invalid_opening_preceding_whitespace = invalid preceding whitespace for frontmatter opening + .note = frontmatter opening should not be preceded by whitespace +parse_frontmatter_length_mismatch = frontmatter close does not match the opening + .label_opening = the opening here has {$len_opening} dashes... + .label_close = ...while the close has {$len_close} dashes +parse_frontmatter_unclosed = unclosed frontmatter + .note = frontmatter opening here was not closed + parse_function_body_equals_expr = function body cannot be `= expression;` .suggestion = surround the expression with `{"{"}` and `{"}"}` instead of `=` and `;` diff --git a/compiler/rustc_parse/src/errors.rs b/compiler/rustc_parse/src/errors.rs index 6a6fb0eb9b5..9e5c81d44a5 100644 --- a/compiler/rustc_parse/src/errors.rs +++ b/compiler/rustc_parse/src/errors.rs @@ -736,6 +736,61 @@ pub(crate) struct FoundExprWouldBeStmt { } #[derive(Diagnostic)] +#[diag(parse_frontmatter_extra_characters_after_close)] +pub(crate) struct FrontmatterExtraCharactersAfterClose { + #[primary_span] + pub span: Span, +} + +#[derive(Diagnostic)] +#[diag(parse_frontmatter_invalid_infostring)] +#[note] +pub(crate) struct FrontmatterInvalidInfostring { + #[primary_span] + pub span: Span, +} + +#[derive(Diagnostic)] +#[diag(parse_frontmatter_invalid_opening_preceding_whitespace)] +pub(crate) struct FrontmatterInvalidOpeningPrecedingWhitespace { + #[primary_span] + pub span: Span, + #[note] + pub note_span: Span, +} + +#[derive(Diagnostic)] +#[diag(parse_frontmatter_unclosed)] +pub(crate) struct FrontmatterUnclosed { + #[primary_span] + pub span: Span, + #[note] + pub note_span: Span, +} + +#[derive(Diagnostic)] +#[diag(parse_frontmatter_invalid_close_preceding_whitespace)] +pub(crate) struct FrontmatterInvalidClosingPrecedingWhitespace { + #[primary_span] + pub span: Span, + #[note] + pub note_span: Span, +} + +#[derive(Diagnostic)] +#[diag(parse_frontmatter_length_mismatch)] +pub(crate) struct FrontmatterLengthMismatch { + #[primary_span] + pub span: Span, + #[label(parse_label_opening)] + pub opening: Span, + #[label(parse_label_close)] + pub close: Span, + pub len_opening: usize, + pub len_close: usize, +} + +#[derive(Diagnostic)] #[diag(parse_leading_plus_not_supported)] pub(crate) struct LeadingPlusNotSupported { #[primary_span] diff --git a/compiler/rustc_parse/src/lexer/mod.rs b/compiler/rustc_parse/src/lexer/mod.rs index e8a5cae54cf..78c5742414b 100644 --- a/compiler/rustc_parse/src/lexer/mod.rs +++ b/compiler/rustc_parse/src/lexer/mod.rs @@ -7,7 +7,9 @@ use rustc_ast::tokenstream::TokenStream; use rustc_ast::util::unicode::contains_text_flow_control_chars; use rustc_errors::codes::*; use rustc_errors::{Applicability, Diag, DiagCtxtHandle, StashKey}; -use rustc_lexer::{Base, Cursor, DocStyle, LiteralKind, RawStrError}; +use rustc_lexer::{ + Base, Cursor, DocStyle, FrontmatterAllowed, LiteralKind, RawStrError, is_whitespace, +}; use rustc_literal_escaper::{EscapeError, Mode, unescape_mixed, unescape_unicode}; use rustc_session::lint::BuiltinLintDiag; use rustc_session::lint::builtin::{ @@ -15,7 +17,7 @@ use rustc_session::lint::builtin::{ TEXT_DIRECTION_CODEPOINT_IN_COMMENT, }; use rustc_session::parse::ParseSess; -use rustc_span::{BytePos, Pos, Span, Symbol}; +use rustc_span::{BytePos, Pos, Span, Symbol, sym}; use tracing::debug; use crate::errors; @@ -56,7 +58,7 @@ pub(crate) fn lex_token_trees<'psess, 'src>( start_pos = start_pos + BytePos::from_usize(shebang_len); } - let cursor = Cursor::new(src); + let cursor = Cursor::new(src, FrontmatterAllowed::Yes); let mut lexer = Lexer { psess, start_pos, @@ -193,6 +195,11 @@ impl<'psess, 'src> Lexer<'psess, 'src> { let content = self.str_from_to(content_start, content_end); self.cook_doc_comment(content_start, content, CommentKind::Block, doc_style) } + rustc_lexer::TokenKind::Frontmatter { has_invalid_preceding_whitespace, invalid_infostring } => { + self.validate_frontmatter(start, has_invalid_preceding_whitespace, invalid_infostring); + preceded_by_whitespace = true; + continue; + } rustc_lexer::TokenKind::Whitespace => { preceded_by_whitespace = true; continue; @@ -256,7 +263,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> { // was consumed. let lit_start = start + BytePos(prefix_len); self.pos = lit_start; - self.cursor = Cursor::new(&str_before[prefix_len as usize..]); + self.cursor = Cursor::new(&str_before[prefix_len as usize..], FrontmatterAllowed::No); self.report_unknown_prefix(start); let prefix_span = self.mk_sp(start, lit_start); return (Token::new(self.ident(start), prefix_span), preceded_by_whitespace); @@ -361,7 +368,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> { // Reset the state so we just lex the `'r`. let lt_start = start + BytePos(2); self.pos = lt_start; - self.cursor = Cursor::new(&str_before[2 as usize..]); + self.cursor = Cursor::new(&str_before[2 as usize..], FrontmatterAllowed::No); let lifetime_name = self.str_from(start); let ident = Symbol::intern(lifetime_name); @@ -474,6 +481,91 @@ impl<'psess, 'src> Lexer<'psess, 'src> { } } + fn validate_frontmatter( + &self, + start: BytePos, + has_invalid_preceding_whitespace: bool, + invalid_infostring: bool, + ) { + let s = self.str_from(start); + let real_start = s.find("---").unwrap(); + let frontmatter_opening_pos = BytePos(real_start as u32) + start; + let s_new = &s[real_start..]; + let within = s_new.trim_start_matches('-'); + let len_opening = s_new.len() - within.len(); + + let frontmatter_opening_end_pos = frontmatter_opening_pos + BytePos(len_opening as u32); + if has_invalid_preceding_whitespace { + let line_start = + BytePos(s[..real_start].rfind("\n").map_or(0, |i| i as u32 + 1)) + start; + let span = self.mk_sp(line_start, frontmatter_opening_end_pos); + let label_span = self.mk_sp(line_start, frontmatter_opening_pos); + self.dcx().emit_err(errors::FrontmatterInvalidOpeningPrecedingWhitespace { + span, + note_span: label_span, + }); + } + + if invalid_infostring { + let line_end = s[real_start..].find('\n').unwrap_or(s[real_start..].len()); + let span = self.mk_sp( + frontmatter_opening_end_pos, + frontmatter_opening_pos + BytePos(line_end as u32), + ); + self.dcx().emit_err(errors::FrontmatterInvalidInfostring { span }); + } + + 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_start_pos = frontmatter_opening_end_pos + BytePos(last_line_start as u32); + + let frontmatter_span = self.mk_sp(frontmatter_opening_pos, self.pos); + self.psess.gated_spans.gate(sym::frontmatter, frontmatter_span); + + if !last_line_trimmed.starts_with("---") { + let label_span = self.mk_sp(frontmatter_opening_pos, frontmatter_opening_end_pos); + self.dcx().emit_err(errors::FrontmatterUnclosed { + span: frontmatter_span, + note_span: label_span, + }); + return; + } + + if last_line_trimmed.len() != last_line.len() { + let line_end = last_line_start_pos + BytePos(last_line.len() as u32); + let span = self.mk_sp(last_line_start_pos, line_end); + let whitespace_end = + last_line_start_pos + BytePos((last_line.len() - last_line_trimmed.len()) as u32); + let label_span = self.mk_sp(last_line_start_pos, whitespace_end); + self.dcx().emit_err(errors::FrontmatterInvalidClosingPrecedingWhitespace { + span, + note_span: label_span, + }); + } + + let rest = last_line_trimmed.trim_start_matches('-'); + let len_close = last_line_trimmed.len() - rest.len(); + if len_close != len_opening { + let span = self.mk_sp(frontmatter_opening_pos, self.pos); + let opening = self.mk_sp(frontmatter_opening_pos, frontmatter_opening_end_pos); + let last_line_close_pos = last_line_start_pos + BytePos(len_close as u32); + let close = self.mk_sp(last_line_start_pos, last_line_close_pos); + self.dcx().emit_err(errors::FrontmatterLengthMismatch { + span, + opening, + close, + len_opening, + len_close, + }); + } + + if !rest.trim_matches(is_whitespace).is_empty() { + let span = self.mk_sp(last_line_start_pos, self.pos); + self.dcx().emit_err(errors::FrontmatterExtraCharactersAfterClose { span }); + } + } + fn cook_doc_comment( &self, content_start: BytePos, @@ -839,7 +931,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> { let space_pos = start + BytePos(1); let space_span = self.mk_sp(space_pos, space_pos); - let mut cursor = Cursor::new(str_before); + let mut cursor = Cursor::new(str_before, FrontmatterAllowed::No); let (is_string, span, unterminated) = match cursor.guarded_double_quoted_string() { Some(rustc_lexer::GuardedStr { n_hashes, terminated, token_len }) => { @@ -905,7 +997,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> { // For backwards compatibility, roll back to after just the first `#` // and return the `Pound` token. self.pos = start + BytePos(1); - self.cursor = Cursor::new(&str_before[1..]); + self.cursor = Cursor::new(&str_before[1..], FrontmatterAllowed::No); token::Pound } } diff --git a/compiler/rustc_session/src/config.rs b/compiler/rustc_session/src/config.rs index a9d9236d318..60e1b465ba9 100644 --- a/compiler/rustc_session/src/config.rs +++ b/compiler/rustc_session/src/config.rs @@ -24,7 +24,8 @@ use rustc_macros::{Decodable, Encodable, HashStable_Generic}; use rustc_span::edition::{DEFAULT_EDITION, EDITION_NAME_LIST, Edition, LATEST_STABLE_EDITION}; use rustc_span::source_map::FilePathMapping; use rustc_span::{ - FileName, FileNameDisplayPreference, RealFileName, SourceFileHashAlgorithm, Symbol, sym, + FileName, FileNameDisplayPreference, FileNameEmbeddablePreference, RealFileName, + SourceFileHashAlgorithm, Symbol, sym, }; use rustc_target::spec::{ FramePointer, LinkSelfContainedComponents, LinkerFeatures, SplitDebuginfo, Target, TargetTuple, @@ -1320,6 +1321,11 @@ fn file_path_mapping( } else { FileNameDisplayPreference::Local }, + if unstable_opts.remap_path_scope.is_all() { + FileNameEmbeddablePreference::RemappedOnly + } else { + FileNameEmbeddablePreference::LocalAndRemapped + }, ) } diff --git a/compiler/rustc_span/src/lib.rs b/compiler/rustc_span/src/lib.rs index 9722031f209..6fcf77e31a2 100644 --- a/compiler/rustc_span/src/lib.rs +++ b/compiler/rustc_span/src/lib.rs @@ -224,7 +224,7 @@ pub fn with_metavar_spans<R>(f: impl FnOnce(&MetavarSpansMap) -> R) -> R { // FIXME: We should use this enum or something like it to get rid of the // use of magic `/rust/1.x/...` paths across the board. -#[derive(Debug, Eq, PartialEq, Clone, Ord, PartialOrd, Decodable)] +#[derive(Debug, Eq, PartialEq, Clone, Ord, PartialOrd, Decodable, Encodable)] pub enum RealFileName { LocalPath(PathBuf), /// For remapped paths (namely paths into libstd that have been mapped @@ -250,28 +250,6 @@ impl Hash for RealFileName { } } -// This is functionally identical to #[derive(Encodable)], with the exception of -// an added assert statement -impl<S: Encoder> Encodable<S> for RealFileName { - fn encode(&self, encoder: &mut S) { - match *self { - RealFileName::LocalPath(ref local_path) => { - encoder.emit_u8(0); - local_path.encode(encoder); - } - - RealFileName::Remapped { ref local_path, ref virtual_name } => { - encoder.emit_u8(1); - // For privacy and build reproducibility, we must not embed host-dependant path - // in artifacts if they have been remapped by --remap-path-prefix - assert!(local_path.is_none()); - local_path.encode(encoder); - virtual_name.encode(encoder); - } - } - } -} - impl RealFileName { /// Returns the path suitable for reading from the file system on the local host, /// if this information exists. @@ -369,6 +347,16 @@ impl From<PathBuf> for FileName { } #[derive(Clone, Copy, Eq, PartialEq, Hash, Debug)] +pub enum FileNameEmbeddablePreference { + /// If a remapped path is available, only embed the `virtual_path` and omit the `local_path`. + /// + /// Otherwise embed the local-path into the `virtual_path`. + RemappedOnly, + /// Embed the original path as well as its remapped `virtual_path` component if available. + LocalAndRemapped, +} + +#[derive(Clone, Copy, Eq, PartialEq, Hash, Debug)] pub enum FileNameDisplayPreference { /// Display the path after the application of rewrite rules provided via `--remap-path-prefix`. /// This is appropriate for paths that get embedded into files produced by the compiler. diff --git a/compiler/rustc_span/src/source_map.rs b/compiler/rustc_span/src/source_map.rs index 0273bb040f4..8a3644163ca 100644 --- a/compiler/rustc_span/src/source_map.rs +++ b/compiler/rustc_span/src/source_map.rs @@ -1108,18 +1108,28 @@ pub fn get_source_map() -> Option<Arc<SourceMap>> { pub struct FilePathMapping { mapping: Vec<(PathBuf, PathBuf)>, filename_display_for_diagnostics: FileNameDisplayPreference, + filename_embeddable_preference: FileNameEmbeddablePreference, } impl FilePathMapping { pub fn empty() -> FilePathMapping { - FilePathMapping::new(Vec::new(), FileNameDisplayPreference::Local) + FilePathMapping::new( + Vec::new(), + FileNameDisplayPreference::Local, + FileNameEmbeddablePreference::RemappedOnly, + ) } pub fn new( mapping: Vec<(PathBuf, PathBuf)>, filename_display_for_diagnostics: FileNameDisplayPreference, + filename_embeddable_preference: FileNameEmbeddablePreference, ) -> FilePathMapping { - FilePathMapping { mapping, filename_display_for_diagnostics } + FilePathMapping { + mapping, + filename_display_for_diagnostics, + filename_embeddable_preference, + } } /// Applies any path prefix substitution as defined by the mapping. @@ -1217,11 +1227,13 @@ impl FilePathMapping { ) -> RealFileName { match file_path { // Anything that's already remapped we don't modify, except for erasing - // the `local_path` portion. - RealFileName::Remapped { local_path: _, virtual_name } => { + // the `local_path` portion (if desired). + RealFileName::Remapped { local_path, virtual_name } => { RealFileName::Remapped { - // We do not want any local path to be exported into metadata - local_path: None, + local_path: match self.filename_embeddable_preference { + FileNameEmbeddablePreference::RemappedOnly => None, + FileNameEmbeddablePreference::LocalAndRemapped => local_path, + }, // We use the remapped name verbatim, even if it looks like a relative // path. The assumption is that the user doesn't want us to further // process paths that have gone through remapping. @@ -1231,12 +1243,18 @@ impl FilePathMapping { RealFileName::LocalPath(unmapped_file_path) => { // If no remapping has been applied yet, try to do so - let (new_path, was_remapped) = self.map_prefix(unmapped_file_path); + let (new_path, was_remapped) = self.map_prefix(&unmapped_file_path); if was_remapped { // It was remapped, so don't modify further return RealFileName::Remapped { - local_path: None, virtual_name: new_path.into_owned(), + // But still provide the local path if desired + local_path: match self.filename_embeddable_preference { + FileNameEmbeddablePreference::RemappedOnly => None, + FileNameEmbeddablePreference::LocalAndRemapped => { + Some(unmapped_file_path) + } + }, }; } @@ -1252,17 +1270,23 @@ impl FilePathMapping { match working_directory { RealFileName::LocalPath(unmapped_working_dir_abs) => { - let file_path_abs = unmapped_working_dir_abs.join(unmapped_file_path_rel); + let unmapped_file_path_abs = + unmapped_working_dir_abs.join(unmapped_file_path_rel); // Although neither `working_directory` nor the file name were subject // to path remapping, the concatenation between the two may be. Hence // we need to do a remapping here. - let (file_path_abs, was_remapped) = self.map_prefix(file_path_abs); + let (file_path_abs, was_remapped) = + self.map_prefix(&unmapped_file_path_abs); if was_remapped { RealFileName::Remapped { - // Erase the actual path - local_path: None, virtual_name: file_path_abs.into_owned(), + local_path: match self.filename_embeddable_preference { + FileNameEmbeddablePreference::RemappedOnly => None, + FileNameEmbeddablePreference::LocalAndRemapped => { + Some(unmapped_file_path_abs) + } + }, } } else { // No kind of remapping applied to this path, so @@ -1271,15 +1295,20 @@ impl FilePathMapping { } } RealFileName::Remapped { - local_path: _, + local_path, virtual_name: remapped_working_dir_abs, } => { // If working_directory has been remapped, then we emit // Remapped variant as the expanded path won't be valid RealFileName::Remapped { - local_path: None, virtual_name: Path::new(remapped_working_dir_abs) - .join(unmapped_file_path_rel), + .join(&unmapped_file_path_rel), + local_path: match self.filename_embeddable_preference { + FileNameEmbeddablePreference::RemappedOnly => None, + FileNameEmbeddablePreference::LocalAndRemapped => local_path + .as_ref() + .map(|local_path| local_path.join(unmapped_file_path_rel)), + }, } } } @@ -1287,27 +1316,6 @@ impl FilePathMapping { } } - /// Expand a relative path to an absolute path **without** remapping taken into account. - /// - /// The resulting `RealFileName` will have its `virtual_path` portion erased if - /// possible (i.e. if there's also a remapped path). - pub fn to_local_embeddable_absolute_path( - &self, - file_path: RealFileName, - working_directory: &RealFileName, - ) -> RealFileName { - let file_path = file_path.local_path_if_available(); - if file_path.is_absolute() { - // No remapping has applied to this path and it is absolute, - // so the working directory cannot influence it either, so - // we are done. - return RealFileName::LocalPath(file_path.to_path_buf()); - } - debug_assert!(file_path.is_relative()); - let working_directory = working_directory.local_path_if_available(); - RealFileName::LocalPath(Path::new(working_directory).join(file_path)) - } - /// Attempts to (heuristically) reverse a prefix mapping. /// /// Returns [`Some`] if there is exactly one mapping where the "to" part is diff --git a/compiler/rustc_span/src/source_map/tests.rs b/compiler/rustc_span/src/source_map/tests.rs index 957f55e3913..589c2a36354 100644 --- a/compiler/rustc_span/src/source_map/tests.rs +++ b/compiler/rustc_span/src/source_map/tests.rs @@ -305,6 +305,7 @@ fn path_prefix_remapping() { let mapping = &FilePathMapping::new( vec![(path("abc/def"), path("foo"))], FileNameDisplayPreference::Remapped, + FileNameEmbeddablePreference::RemappedOnly, ); assert_eq!(map_path_prefix(mapping, "abc/def/src/main.rs"), path_str("foo/src/main.rs")); @@ -316,6 +317,7 @@ fn path_prefix_remapping() { let mapping = &FilePathMapping::new( vec![(path("abc/def"), path("/foo"))], FileNameDisplayPreference::Remapped, + FileNameEmbeddablePreference::RemappedOnly, ); assert_eq!(map_path_prefix(mapping, "abc/def/src/main.rs"), path_str("/foo/src/main.rs")); @@ -327,6 +329,7 @@ fn path_prefix_remapping() { let mapping = &FilePathMapping::new( vec![(path("/abc/def"), path("foo"))], FileNameDisplayPreference::Remapped, + FileNameEmbeddablePreference::RemappedOnly, ); assert_eq!(map_path_prefix(mapping, "/abc/def/src/main.rs"), path_str("foo/src/main.rs")); @@ -338,6 +341,7 @@ fn path_prefix_remapping() { let mapping = &FilePathMapping::new( vec![(path("/abc/def"), path("/foo"))], FileNameDisplayPreference::Remapped, + FileNameEmbeddablePreference::RemappedOnly, ); assert_eq!(map_path_prefix(mapping, "/abc/def/src/main.rs"), path_str("/foo/src/main.rs")); @@ -351,6 +355,7 @@ fn path_prefix_remapping_expand_to_absolute() { let mapping = &FilePathMapping::new( vec![(path("/foo"), path("FOO")), (path("/bar"), path("BAR"))], FileNameDisplayPreference::Remapped, + FileNameEmbeddablePreference::RemappedOnly, ); let working_directory = path("/foo"); let working_directory = RealFileName::Remapped { @@ -449,12 +454,78 @@ fn path_prefix_remapping_expand_to_absolute() { } #[test] +fn path_prefix_remapping_expand_to_absolute_and_local() { + // "virtual" working directory is relative path + let mapping = &FilePathMapping::new( + vec![(path("/foo"), path("FOO")), (path("/bar"), path("BAR"))], + FileNameDisplayPreference::Remapped, + FileNameEmbeddablePreference::LocalAndRemapped, + ); + let working_directory = path("/foo"); + let working_directory = RealFileName::Remapped { + local_path: Some(working_directory.clone()), + virtual_name: mapping.map_prefix(working_directory).0.into_owned(), + }; + + assert_eq!(working_directory.remapped_path_if_available(), path("FOO")); + + // Unmapped absolute path + assert_eq!( + mapping.to_embeddable_absolute_path( + RealFileName::LocalPath(path("/foo/src/main.rs")), + &working_directory + ), + RealFileName::Remapped { + local_path: Some(path("/foo/src/main.rs")), + virtual_name: path("FOO/src/main.rs") + } + ); + + // Unmapped absolute path with unrelated working directory + assert_eq!( + mapping.to_embeddable_absolute_path( + RealFileName::LocalPath(path("/bar/src/main.rs")), + &working_directory + ), + RealFileName::Remapped { + local_path: Some(path("/bar/src/main.rs")), + virtual_name: path("BAR/src/main.rs") + } + ); + + // Already remapped absolute path, with unrelated working directory + assert_eq!( + mapping.to_embeddable_absolute_path( + RealFileName::Remapped { + local_path: Some(path("/bar/src/main.rs")), + virtual_name: path("BAR/src/main.rs"), + }, + &working_directory + ), + RealFileName::Remapped { + local_path: Some(path("/bar/src/main.rs")), + virtual_name: path("BAR/src/main.rs") + } + ); + + // Already remapped relative path + assert_eq!( + mapping.to_embeddable_absolute_path( + RealFileName::Remapped { local_path: None, virtual_name: path("XYZ/src/main.rs") }, + &working_directory + ), + RealFileName::Remapped { local_path: None, virtual_name: path("XYZ/src/main.rs") } + ); +} + +#[test] fn path_prefix_remapping_reverse() { // Ignores options without alphanumeric chars. { let mapping = &FilePathMapping::new( vec![(path("abc"), path("/")), (path("def"), path("."))], FileNameDisplayPreference::Remapped, + FileNameEmbeddablePreference::RemappedOnly, ); assert_eq!(reverse_map_prefix(mapping, "/hello.rs"), None); @@ -466,6 +537,7 @@ fn path_prefix_remapping_reverse() { let mapping = &FilePathMapping::new( vec![(path("abc"), path("/redacted")), (path("def"), path("/redacted"))], FileNameDisplayPreference::Remapped, + FileNameEmbeddablePreference::RemappedOnly, ); assert_eq!(reverse_map_prefix(mapping, "/redacted/hello.rs"), None); @@ -476,6 +548,7 @@ fn path_prefix_remapping_reverse() { let mapping = &FilePathMapping::new( vec![(path("abc"), path("/redacted")), (path("def/ghi"), path("/fake/dir"))], FileNameDisplayPreference::Remapped, + FileNameEmbeddablePreference::RemappedOnly, ); assert_eq!( diff --git a/compiler/rustc_span/src/symbol.rs b/compiler/rustc_span/src/symbol.rs index 31d129c3465..ea142e3d482 100644 --- a/compiler/rustc_span/src/symbol.rs +++ b/compiler/rustc_span/src/symbol.rs @@ -1049,6 +1049,7 @@ symbols! { from_u16, from_usize, from_yeet, + frontmatter, fs_create_dir, fsub_algebraic, fsub_fast, diff --git a/compiler/rustc_target/src/spec/mod.rs b/compiler/rustc_target/src/spec/mod.rs index 37ea0d6e7b5..303be54a6d7 100644 --- a/compiler/rustc_target/src/spec/mod.rs +++ b/compiler/rustc_target/src/spec/mod.rs @@ -3567,7 +3567,19 @@ impl Target { "x86" => (Architecture::I386, None), "s390x" => (Architecture::S390x, None), "mips" | "mips32r6" => (Architecture::Mips, None), - "mips64" | "mips64r6" => (Architecture::Mips64, None), + "mips64" | "mips64r6" => ( + // While there are currently no builtin targets + // using the N32 ABI, it is possible to specify + // it using a custom target specification. N32 + // is an ILP32 ABI like the Aarch64_Ilp32 + // and X86_64_X32 cases above and below this one. + if self.options.llvm_abiname.as_ref() == "n32" { + Architecture::Mips64_N32 + } else { + Architecture::Mips64 + }, + None, + ), "x86_64" => ( if self.pointer_width == 32 { Architecture::X86_64_X32 diff --git a/compiler/rustc_target/src/spec/targets/mips64_openwrt_linux_musl.rs b/compiler/rustc_target/src/spec/targets/mips64_openwrt_linux_musl.rs index 71b3fbe00b2..508abc01018 100644 --- a/compiler/rustc_target/src/spec/targets/mips64_openwrt_linux_musl.rs +++ b/compiler/rustc_target/src/spec/targets/mips64_openwrt_linux_musl.rs @@ -27,6 +27,7 @@ pub(crate) fn target() -> Target { abi: "abi64".into(), endian: Endian::Big, mcount: "_mcount".into(), + llvm_abiname: "n64".into(), ..base }, } diff --git a/compiler/rustc_target/src/spec/targets/mips64_unknown_linux_gnuabi64.rs b/compiler/rustc_target/src/spec/targets/mips64_unknown_linux_gnuabi64.rs index b130ca29c7f..a26350ff225 100644 --- a/compiler/rustc_target/src/spec/targets/mips64_unknown_linux_gnuabi64.rs +++ b/compiler/rustc_target/src/spec/targets/mips64_unknown_linux_gnuabi64.rs @@ -22,6 +22,7 @@ pub(crate) fn target() -> Target { features: "+mips64r2,+xgot".into(), max_atomic_width: Some(64), mcount: "_mcount".into(), + llvm_abiname: "n64".into(), ..base::linux_gnu::opts() }, diff --git a/compiler/rustc_target/src/spec/targets/mips64_unknown_linux_muslabi64.rs b/compiler/rustc_target/src/spec/targets/mips64_unknown_linux_muslabi64.rs index 4ea7c7bff44..fd509503053 100644 --- a/compiler/rustc_target/src/spec/targets/mips64_unknown_linux_muslabi64.rs +++ b/compiler/rustc_target/src/spec/targets/mips64_unknown_linux_muslabi64.rs @@ -25,6 +25,7 @@ pub(crate) fn target() -> Target { mcount: "_mcount".into(), // FIXME(compiler-team#422): musl targets should be dynamically linked by default. crt_static_default: true, + llvm_abiname: "n64".into(), ..base }, } diff --git a/compiler/rustc_target/src/spec/targets/mips64el_unknown_linux_gnuabi64.rs b/compiler/rustc_target/src/spec/targets/mips64el_unknown_linux_gnuabi64.rs index a9afea27ef3..19bceadc622 100644 --- a/compiler/rustc_target/src/spec/targets/mips64el_unknown_linux_gnuabi64.rs +++ b/compiler/rustc_target/src/spec/targets/mips64el_unknown_linux_gnuabi64.rs @@ -19,6 +19,7 @@ pub(crate) fn target() -> Target { features: "+mips64r2,+xgot".into(), max_atomic_width: Some(64), mcount: "_mcount".into(), + llvm_abiname: "n64".into(), ..base::linux_gnu::opts() }, diff --git a/compiler/rustc_target/src/spec/targets/mips64el_unknown_linux_muslabi64.rs b/compiler/rustc_target/src/spec/targets/mips64el_unknown_linux_muslabi64.rs index 7bdd9edda70..aa087b1a35a 100644 --- a/compiler/rustc_target/src/spec/targets/mips64el_unknown_linux_muslabi64.rs +++ b/compiler/rustc_target/src/spec/targets/mips64el_unknown_linux_muslabi64.rs @@ -19,6 +19,11 @@ pub(crate) fn target() -> Target { pointer_width: 64, data_layout: "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128".into(), arch: "mips64".into(), - options: TargetOptions { abi: "abi64".into(), mcount: "_mcount".into(), ..base }, + options: TargetOptions { + abi: "abi64".into(), + mcount: "_mcount".into(), + llvm_abiname: "n64".into(), + ..base + }, } } diff --git a/compiler/rustc_target/src/spec/targets/mipsisa64r6_unknown_linux_gnuabi64.rs b/compiler/rustc_target/src/spec/targets/mipsisa64r6_unknown_linux_gnuabi64.rs index 3eefa27ea04..cdd5f6b8436 100644 --- a/compiler/rustc_target/src/spec/targets/mipsisa64r6_unknown_linux_gnuabi64.rs +++ b/compiler/rustc_target/src/spec/targets/mipsisa64r6_unknown_linux_gnuabi64.rs @@ -22,6 +22,7 @@ pub(crate) fn target() -> Target { features: "+mips64r6".into(), max_atomic_width: Some(64), mcount: "_mcount".into(), + llvm_abiname: "n64".into(), ..base::linux_gnu::opts() }, diff --git a/compiler/rustc_target/src/spec/targets/mipsisa64r6el_unknown_linux_gnuabi64.rs b/compiler/rustc_target/src/spec/targets/mipsisa64r6el_unknown_linux_gnuabi64.rs index 0887180791c..88879a25818 100644 --- a/compiler/rustc_target/src/spec/targets/mipsisa64r6el_unknown_linux_gnuabi64.rs +++ b/compiler/rustc_target/src/spec/targets/mipsisa64r6el_unknown_linux_gnuabi64.rs @@ -19,6 +19,7 @@ pub(crate) fn target() -> Target { features: "+mips64r6".into(), max_atomic_width: Some(64), mcount: "_mcount".into(), + llvm_abiname: "n64".into(), ..base::linux_gnu::opts() }, diff --git a/compiler/stable_mir/Cargo.toml b/compiler/stable_mir/Cargo.toml index 3a01ee5783e..516c8e9c718 100644 --- a/compiler/stable_mir/Cargo.toml +++ b/compiler/stable_mir/Cargo.toml @@ -5,3 +5,9 @@ edition = "2024" [dependencies] rustc_smir = { path = "../rustc_smir" } + +[features] +# Provides access to APIs that expose internals of the rust compiler. +# APIs enabled by this feature are unstable. They can be removed or modified +# at any point and they are not included in the crate's semantic versioning. +rustc_internal = [] diff --git a/compiler/stable_mir/src/lib.rs b/compiler/stable_mir/src/lib.rs index cc0fb52433d..688f3936b26 100644 --- a/compiler/stable_mir/src/lib.rs +++ b/compiler/stable_mir/src/lib.rs @@ -4,4 +4,8 @@ //! This is a transitional measure as described in [PR #139319](https://github.com/rust-lang/rust/pull/139319). //! Once the refactoring is complete, the `stable_mir` implementation will be moved back here. +/// Export the rustc_internal APIs. Note that this module has no stability +/// guarantees and it is not taken into account for semver. +#[cfg(feature = "rustc_internal")] +pub use rustc_smir::rustc_internal; pub use rustc_smir::stable_mir::*; |
