diff options
95 files changed, 960 insertions, 1229 deletions
diff --git a/Cargo.lock b/Cargo.lock index 99bd1bff020..52c9269fa87 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -239,9 +239,9 @@ checksum = "5c8214115b7bf84099f1309324e63141d4c5d7cc26862f97a0a857dbefe165bd" [[package]] name = "blake3" -version = "1.8.0" +version = "1.8.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "34a796731680be7931955498a16a10b2270c7762963d5d570fdbfe02dcbf314f" +checksum = "389a099b34312839e16420d499a9cad9650541715937ffbdd40d36f49e77eeb3" dependencies = [ "arrayref", "arrayvec", @@ -800,9 +800,9 @@ dependencies = [ [[package]] name = "ctrlc" -version = "3.4.5" +version = "3.4.6" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "90eeab0aa92f3f9b4e87f258c72b139c207d251f9cbc1080a0086b86a8870dd3" +checksum = "697b5419f348fd5ae2478e8018cb016c00a5881c7f46c717de98ffd135a5651c" dependencies = [ "nix", "windows-sys 0.59.0", @@ -1079,9 +1079,9 @@ dependencies = [ [[package]] name = "env_logger" -version = "0.11.7" +version = "0.11.8" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c3716d7a920fb4fac5d84e9d4bce8ceb321e9414b4409da61b07b75c1e3d0697" +checksum = "13c863f0904021b108aa8b2f55046443e6b1ebde8fd4a15c399893aae4fa069f" dependencies = [ "anstream", "anstyle", @@ -1098,9 +1098,9 @@ checksum = "877a4ace8713b0bcf2a4e7eec82529c029f1d0619886d18145fea96c3ffe5c0f" [[package]] name = "errno" -version = "0.3.10" +version = "0.3.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "33d852cb9b869c2a9b3df2f71a3074817f01e1844f839a144f5fcef059a4eb5d" +checksum = "976dd42dc7e85965fe702eb8164f21f450704bdde31faefd6471dba214cb594e" dependencies = [ "libc", "windows-sys 0.59.0", @@ -1163,12 +1163,12 @@ dependencies = [ [[package]] name = "flate2" -version = "1.1.0" +version = "1.1.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "11faaf5a5236997af9848be0bef4db95824b1d534ebc64d0f0c6cf3e67bd38dc" +checksum = "7ced92e76e966ca2fd84c8f7aa01a4aea65b0eb6648d72f7c8f3e2764a67fece" dependencies = [ "crc32fast", - "miniz_oxide 0.8.5", + "miniz_oxide 0.8.7", ] [[package]] @@ -1773,9 +1773,9 @@ checksum = "ce23b50ad8242c51a442f3ff322d56b02f08852c77e4c0b4d3fd684abc89c683" [[package]] name = "indexmap" -version = "2.8.0" +version = "2.9.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3954d50fe15b02142bf25d3b8bdadb634ec3948f103d04ffe3031bc8fe9d7058" +checksum = "cea70ddb795996207ad57735b50c5982d8844f38ba9ee5f1aedcfb708a2aa11e" dependencies = [ "equivalent", "hashbrown", @@ -2252,9 +2252,9 @@ dependencies = [ [[package]] name = "miniz_oxide" -version = "0.8.5" +version = "0.8.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8e3e04debbb59698c15bacbb6d93584a8c0ca9cc3213cb423d31f760d8843ce5" +checksum = "ff70ce3e48ae43fa075863cef62e8b43b71a4f2382229920e0df362592919430" dependencies = [ "adler2", ] @@ -2506,9 +2506,9 @@ checksum = "d05e27ee213611ffe7d6348b942e8f942b37114c00cc03cec254295a4a17852e" [[package]] name = "openssl-sys" -version = "0.9.106" +version = "0.9.107" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8bb61ea9811cc39e3c2069f40b8b8e2e70d8569b361f879786cc7ed48b777cdd" +checksum = "8288979acd84749c744a9014b4382d42b8f7b2592847b5afb2ed29e5d16ede07" dependencies = [ "cc", "libc", @@ -2970,9 +2970,9 @@ dependencies = [ [[package]] name = "redox_syscall" -version = "0.5.10" +version = "0.5.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0b8c0c260b63a8219631167be35e6a988e9554dbd323f8bd08439c8ed1302bd1" +checksum = "d2f103c6d277498fbceb16e84d317e2a400f160f46904d5f5410848c829511a3" dependencies = [ "bitflags", ] @@ -3151,6 +3151,12 @@ source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "357703d41365b4b27c590e3ed91eabb1b663f07c4c084095e60cbed4362dff0d" [[package]] +name = "rustc-literal-escaper" +version = "0.0.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0041b6238913c41fe704213a4a9329e2f685a156d1781998128b4149c230ad04" + +[[package]] name = "rustc-main" version = "0.0.0" dependencies = [ @@ -3242,10 +3248,10 @@ version = "0.0.0" dependencies = [ "bitflags", "memchr", + "rustc-literal-escaper", "rustc_ast_ir", "rustc_data_structures", "rustc_index", - "rustc_lexer", "rustc_macros", "rustc_serialize", "rustc_span", @@ -4200,6 +4206,7 @@ name = "rustc_parse" version = "0.0.0" dependencies = [ "bitflags", + "rustc-literal-escaper", "rustc_ast", "rustc_ast_pretty", "rustc_data_structures", @@ -4222,6 +4229,7 @@ dependencies = [ name = "rustc_parse_format" version = "0.0.0" dependencies = [ + "rustc-literal-escaper", "rustc_index", "rustc_lexer", ] @@ -4581,6 +4589,7 @@ version = "0.0.0" dependencies = [ "bitflags", "derive-where", + "ena", "indexmap", "rustc-hash 1.1.0", "rustc_ast_ir", @@ -4918,9 +4927,9 @@ dependencies = [ [[package]] name = "smallvec" -version = "1.14.0" +version = "1.15.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7fcf8323ef1faaee30a44a340193b1ac6814fd9b7b4e88e9d4519a3e4abe1cfd" +checksum = "8917285742e9f3e1683f0a9c4e6b57960b7314d0b08d30d1ecd426713ee2eee9" [[package]] name = "socket2" @@ -5368,9 +5377,9 @@ checksum = "1f3ccbac311fea05f86f61904b462b55fb3df8837a366dfc601a0161d0532f20" [[package]] name = "tokio" -version = "1.44.1" +version = "1.44.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f382da615b842244d4b8738c82ed1275e6c5dd90c459a30941cd07080b06c91a" +checksum = "e6b88822cbe49de4185e3a4cbf8321dd487cf5fe0c5c65695fef6346371e9c48" dependencies = [ "backtrace", "bytes", diff --git a/compiler/rustc_ast/Cargo.toml b/compiler/rustc_ast/Cargo.toml index 902287d0328..b2d3b90fc44 100644 --- a/compiler/rustc_ast/Cargo.toml +++ b/compiler/rustc_ast/Cargo.toml @@ -7,10 +7,10 @@ edition = "2024" # tidy-alphabetical-start bitflags = "2.4.1" memchr = "2.7.4" +rustc-literal-escaper = "0.0.2" rustc_ast_ir = { path = "../rustc_ast_ir" } rustc_data_structures = { path = "../rustc_data_structures" } rustc_index = { path = "../rustc_index" } -rustc_lexer = { path = "../rustc_lexer" } rustc_macros = { path = "../rustc_macros" } rustc_serialize = { path = "../rustc_serialize" } rustc_span = { path = "../rustc_span" } diff --git a/compiler/rustc_ast/src/ast.rs b/compiler/rustc_ast/src/ast.rs index 97e6879c33e..6362b539fc7 100644 --- a/compiler/rustc_ast/src/ast.rs +++ b/compiler/rustc_ast/src/ast.rs @@ -563,6 +563,7 @@ impl Pat { /// This is intended for use by diagnostics. pub fn to_ty(&self) -> Option<P<Ty>> { let kind = match &self.kind { + PatKind::Missing => unreachable!(), // In a type expression `_` is an inference variable. PatKind::Wild => TyKind::Infer, // An IDENT pattern with no binding mode would be valid as path to a type. E.g. `u32`. @@ -625,7 +626,8 @@ impl Pat { | PatKind::Guard(s, _) => s.walk(it), // These patterns do not contain subpatterns, skip. - PatKind::Wild + PatKind::Missing + | PatKind::Wild | PatKind::Rest | PatKind::Never | PatKind::Expr(_) @@ -676,6 +678,7 @@ impl Pat { /// Return a name suitable for diagnostics. pub fn descr(&self) -> Option<String> { match &self.kind { + PatKind::Missing => unreachable!(), PatKind::Wild => Some("_".to_string()), PatKind::Ident(BindingMode::NONE, ident, None) => Some(format!("{ident}")), PatKind::Ref(pat, mutbl) => pat.descr().map(|d| format!("&{}{d}", mutbl.prefix_str())), @@ -769,6 +772,9 @@ pub enum RangeSyntax { // Adding a new variant? Please update `test_pat` in `tests/ui/macros/stringify.rs`. #[derive(Clone, Encodable, Decodable, Debug)] pub enum PatKind { + /// A missing pattern, e.g. for an anonymous param in a bare fn like `fn f(u32)`. + Missing, + /// Represents a wildcard pattern (`_`). Wild, diff --git a/compiler/rustc_ast/src/mut_visit.rs b/compiler/rustc_ast/src/mut_visit.rs index f7d13acdfc4..00934b73a88 100644 --- a/compiler/rustc_ast/src/mut_visit.rs +++ b/compiler/rustc_ast/src/mut_visit.rs @@ -1587,7 +1587,7 @@ pub fn walk_pat<T: MutVisitor>(vis: &mut T, pat: &mut P<Pat>) { vis.visit_id(id); match kind { PatKind::Err(_guar) => {} - PatKind::Wild | PatKind::Rest | PatKind::Never => {} + PatKind::Missing | PatKind::Wild | PatKind::Rest | PatKind::Never => {} PatKind::Ident(_binding_mode, ident, sub) => { vis.visit_ident(ident); visit_opt(sub, |sub| vis.visit_pat(sub)); diff --git a/compiler/rustc_ast/src/util/literal.rs b/compiler/rustc_ast/src/util/literal.rs index 6896ac723fa..b8526cf9d95 100644 --- a/compiler/rustc_ast/src/util/literal.rs +++ b/compiler/rustc_ast/src/util/literal.rs @@ -2,7 +2,7 @@ use std::{ascii, fmt, str}; -use rustc_lexer::unescape::{ +use rustc_literal_escaper::{ MixedUnit, Mode, byte_from_char, unescape_byte, unescape_char, unescape_mixed, unescape_unicode, }; use rustc_span::{Span, Symbol, kw, sym}; diff --git a/compiler/rustc_ast/src/visit.rs b/compiler/rustc_ast/src/visit.rs index 1ef92ff8898..ac6aa6b9e72 100644 --- a/compiler/rustc_ast/src/visit.rs +++ b/compiler/rustc_ast/src/visit.rs @@ -750,7 +750,7 @@ pub fn walk_pat<'a, V: Visitor<'a>>(visitor: &mut V, pattern: &'a Pat) -> V::Res try_visit!(visitor.visit_pat(subpattern)); try_visit!(visitor.visit_expr(guard_condition)); } - PatKind::Wild | PatKind::Rest | PatKind::Never => {} + PatKind::Missing | PatKind::Wild | PatKind::Rest | PatKind::Never => {} PatKind::Err(_guar) => {} PatKind::Tuple(elems) | PatKind::Slice(elems) | PatKind::Or(elems) => { walk_list!(visitor, visit_pat, elems); diff --git a/compiler/rustc_ast_lowering/src/expr.rs b/compiler/rustc_ast_lowering/src/expr.rs index 80bb1e8fc41..df4fd2ef52f 100644 --- a/compiler/rustc_ast_lowering/src/expr.rs +++ b/compiler/rustc_ast_lowering/src/expr.rs @@ -74,14 +74,16 @@ impl<'hir> LoweringContext<'_, 'hir> { // Merge attributes into the inner expression. if !e.attrs.is_empty() { let old_attrs = self.attrs.get(&ex.hir_id.local_id).copied().unwrap_or(&[]); - self.attrs.insert( - ex.hir_id.local_id, - &*self.arena.alloc_from_iter( - self.lower_attrs_vec(&e.attrs, e.span) - .into_iter() - .chain(old_attrs.iter().cloned()), - ), + let attrs = &*self.arena.alloc_from_iter( + self.lower_attrs_vec(&e.attrs, e.span) + .into_iter() + .chain(old_attrs.iter().cloned()), ); + if attrs.is_empty() { + return ex; + } + + self.attrs.insert(ex.hir_id.local_id, attrs); } return ex; } diff --git a/compiler/rustc_ast_lowering/src/lib.rs b/compiler/rustc_ast_lowering/src/lib.rs index d5d6dcd8d63..38190d44fc5 100644 --- a/compiler/rustc_ast_lowering/src/lib.rs +++ b/compiler/rustc_ast_lowering/src/lib.rs @@ -1496,18 +1496,13 @@ impl<'a, 'hir> LoweringContext<'a, 'hir> { fn lower_fn_params_to_names(&mut self, decl: &FnDecl) -> &'hir [Option<Ident>] { self.arena.alloc_from_iter(decl.inputs.iter().map(|param| match param.pat.kind { - PatKind::Ident(_, ident, _) => { - if ident.name != kw::Empty { - Some(self.lower_ident(ident)) - } else { - None - } - } + PatKind::Missing => None, + PatKind::Ident(_, ident, _) => Some(self.lower_ident(ident)), PatKind::Wild => Some(Ident::new(kw::Underscore, self.lower_span(param.pat.span))), _ => { self.dcx().span_delayed_bug( param.pat.span, - "non-ident/wild param pat must trigger an error", + "non-missing/ident/wild param pat must trigger an error", ); None } diff --git a/compiler/rustc_ast_lowering/src/pat.rs b/compiler/rustc_ast_lowering/src/pat.rs index 07cc64a1358..f94d788a9b0 100644 --- a/compiler/rustc_ast_lowering/src/pat.rs +++ b/compiler/rustc_ast_lowering/src/pat.rs @@ -26,6 +26,7 @@ impl<'a, 'hir> LoweringContext<'a, 'hir> { let pat_hir_id = self.lower_node_id(pattern.id); let node = loop { match &pattern.kind { + PatKind::Missing => break hir::PatKind::Missing, PatKind::Wild => break hir::PatKind::Wild, PatKind::Never => break hir::PatKind::Never, PatKind::Ident(binding_mode, ident, sub) => { diff --git a/compiler/rustc_ast_passes/src/ast_validation.rs b/compiler/rustc_ast_passes/src/ast_validation.rs index 9916de8b7b1..dc77e7b2834 100644 --- a/compiler/rustc_ast_passes/src/ast_validation.rs +++ b/compiler/rustc_ast_passes/src/ast_validation.rs @@ -244,7 +244,7 @@ impl<'a> AstValidator<'a> { fn check_decl_no_pat(decl: &FnDecl, mut report_err: impl FnMut(Span, Option<Ident>, bool)) { for Param { pat, .. } in &decl.inputs { match pat.kind { - PatKind::Ident(BindingMode::NONE, _, None) | PatKind::Wild => {} + PatKind::Missing | PatKind::Ident(BindingMode::NONE, _, None) | PatKind::Wild => {} PatKind::Ident(BindingMode::MUT, ident, None) => { report_err(pat.span, Some(ident), true) } diff --git a/compiler/rustc_ast_pretty/src/pprust/state.rs b/compiler/rustc_ast_pretty/src/pprust/state.rs index 3dbfc191f8f..683ad5c66bf 100644 --- a/compiler/rustc_ast_pretty/src/pprust/state.rs +++ b/compiler/rustc_ast_pretty/src/pprust/state.rs @@ -1622,9 +1622,9 @@ impl<'a> State<'a> { fn print_pat(&mut self, pat: &ast::Pat) { self.maybe_print_comment(pat.span.lo()); self.ann.pre(self, AnnNode::Pat(pat)); - /* Pat isn't normalized, but the beauty of it - is that it doesn't matter */ + /* Pat isn't normalized, but the beauty of it is that it doesn't matter */ match &pat.kind { + PatKind::Missing => unreachable!(), PatKind::Wild => self.word("_"), PatKind::Never => self.word("!"), PatKind::Ident(BindingMode(by_ref, mutbl), ident, sub) => { @@ -1946,12 +1946,7 @@ impl<'a> State<'a> { if let Some(eself) = input.to_self() { self.print_explicit_self(&eself); } else { - let invalid = if let PatKind::Ident(_, ident, _) = input.pat.kind { - ident.name == kw::Empty - } else { - false - }; - if !invalid { + if !matches!(input.pat.kind, PatKind::Missing) { self.print_pat(&input.pat); self.word(":"); self.space(); diff --git a/compiler/rustc_builtin_macros/src/autodiff.rs b/compiler/rustc_builtin_macros/src/autodiff.rs index 7f99f75b2b9..351413dea49 100644 --- a/compiler/rustc_builtin_macros/src/autodiff.rs +++ b/compiler/rustc_builtin_macros/src/autodiff.rs @@ -17,7 +17,7 @@ mod llvm_enzyme { use rustc_ast::visit::AssocCtxt::*; use rustc_ast::{ self as ast, AssocItemKind, BindingMode, ExprKind, FnRetTy, FnSig, Generics, ItemKind, - MetaItemInner, PatKind, QSelf, TyKind, + MetaItemInner, PatKind, QSelf, TyKind, Visibility, }; use rustc_expand::base::{Annotatable, ExtCtxt}; use rustc_span::{Ident, Span, Symbol, kw, sym}; @@ -72,6 +72,16 @@ mod llvm_enzyme { } } + // Get information about the function the macro is applied to + fn extract_item_info(iitem: &P<ast::Item>) -> Option<(Visibility, FnSig, Ident)> { + match &iitem.kind { + ItemKind::Fn(box ast::Fn { sig, ident, .. }) => { + Some((iitem.vis.clone(), sig.clone(), ident.clone())) + } + _ => None, + } + } + pub(crate) fn from_ast( ecx: &mut ExtCtxt<'_>, meta_item: &ThinVec<MetaItemInner>, @@ -199,32 +209,26 @@ mod llvm_enzyme { return vec![item]; } let dcx = ecx.sess.dcx(); - // first get the annotable item: - let (primal, sig, is_impl): (Ident, FnSig, bool) = match &item { - Annotatable::Item(iitem) => { - let (ident, sig) = match &iitem.kind { - ItemKind::Fn(box ast::Fn { ident, sig, .. }) => (ident, sig), - _ => { - dcx.emit_err(errors::AutoDiffInvalidApplication { span: item.span() }); - return vec![item]; - } - }; - (*ident, sig.clone(), false) - } + + // first get information about the annotable item: + let Some((vis, sig, primal)) = (match &item { + Annotatable::Item(iitem) => extract_item_info(iitem), + Annotatable::Stmt(stmt) => match &stmt.kind { + ast::StmtKind::Item(iitem) => extract_item_info(iitem), + _ => None, + }, Annotatable::AssocItem(assoc_item, Impl { of_trait: false }) => { - let (ident, sig) = match &assoc_item.kind { - ast::AssocItemKind::Fn(box ast::Fn { ident, sig, .. }) => (ident, sig), - _ => { - dcx.emit_err(errors::AutoDiffInvalidApplication { span: item.span() }); - return vec![item]; + match &assoc_item.kind { + ast::AssocItemKind::Fn(box ast::Fn { sig, ident, .. }) => { + Some((assoc_item.vis.clone(), sig.clone(), ident.clone())) } - }; - (*ident, sig.clone(), true) - } - _ => { - dcx.emit_err(errors::AutoDiffInvalidApplication { span: item.span() }); - return vec![item]; + _ => None, + } } + _ => None, + }) else { + dcx.emit_err(errors::AutoDiffInvalidApplication { span: item.span() }); + return vec![item]; }; let meta_item_vec: ThinVec<MetaItemInner> = match meta_item.kind { @@ -238,15 +242,6 @@ mod llvm_enzyme { let has_ret = has_ret(&sig.decl.output); let sig_span = ecx.with_call_site_ctxt(sig.span); - let vis = match &item { - Annotatable::Item(iitem) => iitem.vis.clone(), - Annotatable::AssocItem(assoc_item, _) => assoc_item.vis.clone(), - _ => { - dcx.emit_err(errors::AutoDiffInvalidApplication { span: item.span() }); - return vec![item]; - } - }; - // create TokenStream from vec elemtents: // meta_item doesn't have a .tokens field let mut ts: Vec<TokenTree> = vec![]; @@ -379,6 +374,22 @@ mod llvm_enzyme { } Annotatable::AssocItem(assoc_item.clone(), i) } + Annotatable::Stmt(ref mut stmt) => { + match stmt.kind { + ast::StmtKind::Item(ref mut iitem) => { + if !iitem.attrs.iter().any(|a| same_attribute(&a.kind, &attr.kind)) { + iitem.attrs.push(attr); + } + if !iitem.attrs.iter().any(|a| same_attribute(&a.kind, &inline_never.kind)) + { + iitem.attrs.push(inline_never.clone()); + } + } + _ => unreachable!("stmt kind checked previously"), + }; + + Annotatable::Stmt(stmt.clone()) + } _ => { unreachable!("annotatable kind checked previously") } @@ -389,22 +400,40 @@ mod llvm_enzyme { delim: rustc_ast::token::Delimiter::Parenthesis, tokens: ts, }); + let d_attr = outer_normal_attr(&rustc_ad_attr, new_id, span); - let d_annotatable = if is_impl { - let assoc_item: AssocItemKind = ast::AssocItemKind::Fn(asdf); - let d_fn = P(ast::AssocItem { - attrs: thin_vec![d_attr, inline_never], - id: ast::DUMMY_NODE_ID, - span, - vis, - kind: assoc_item, - tokens: None, - }); - Annotatable::AssocItem(d_fn, Impl { of_trait: false }) - } else { - let mut d_fn = ecx.item(span, thin_vec![d_attr, inline_never], ItemKind::Fn(asdf)); - d_fn.vis = vis; - Annotatable::Item(d_fn) + let d_annotatable = match &item { + Annotatable::AssocItem(_, _) => { + let assoc_item: AssocItemKind = ast::AssocItemKind::Fn(asdf); + let d_fn = P(ast::AssocItem { + attrs: thin_vec![d_attr, inline_never], + id: ast::DUMMY_NODE_ID, + span, + vis, + kind: assoc_item, + tokens: None, + }); + Annotatable::AssocItem(d_fn, Impl { of_trait: false }) + } + Annotatable::Item(_) => { + let mut d_fn = ecx.item(span, thin_vec![d_attr, inline_never], ItemKind::Fn(asdf)); + d_fn.vis = vis; + + Annotatable::Item(d_fn) + } + Annotatable::Stmt(_) => { + let mut d_fn = ecx.item(span, thin_vec![d_attr, inline_never], ItemKind::Fn(asdf)); + d_fn.vis = vis; + + Annotatable::Stmt(P(ast::Stmt { + id: ast::DUMMY_NODE_ID, + kind: ast::StmtKind::Item(d_fn), + span, + })) + } + _ => { + unreachable!("item kind checked previously") + } }; return vec![orig_annotatable, d_annotatable]; diff --git a/compiler/rustc_codegen_llvm/src/back/write.rs b/compiler/rustc_codegen_llvm/src/back/write.rs index bf6138142b6..76d431a4975 100644 --- a/compiler/rustc_codegen_llvm/src/back/write.rs +++ b/compiler/rustc_codegen_llvm/src/back/write.rs @@ -439,12 +439,9 @@ fn report_inline_asm( let span = if cookie == 0 || matches!(cgcx.lto, Lto::Fat | Lto::Thin) { SpanData::default() } else { - let lo = BytePos::from_u32(cookie as u32); - let hi = BytePos::from_u32((cookie >> 32) as u32); SpanData { - lo, - // LLVM version < 19 silently truncates the cookie to 32 bits in some situations. - hi: if hi.to_u32() != 0 { hi } else { lo }, + lo: BytePos::from_u32(cookie as u32), + hi: BytePos::from_u32((cookie >> 32) as u32), ctxt: SyntaxContext::root(), parent: None, } diff --git a/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen.rs b/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen.rs index 9a2473d6cf2..5e62ce285dd 100644 --- a/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen.rs +++ b/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen.rs @@ -5,15 +5,11 @@ use rustc_abi::Align; use rustc_codegen_ssa::traits::{ BaseTypeCodegenMethods, ConstCodegenMethods, StaticCodegenMethods, }; -use rustc_data_structures::fx::{FxHashSet, FxIndexMap}; -use rustc_hir::def_id::{DefId, LocalDefId}; +use rustc_data_structures::fx::FxIndexMap; use rustc_index::IndexVec; -use rustc_middle::mir; -use rustc_middle::mir::mono::MonoItemPartitions; -use rustc_middle::ty::{self, TyCtxt}; +use rustc_middle::ty::TyCtxt; use rustc_session::RemapFileNameExt; use rustc_session::config::RemapPathScopeComponents; -use rustc_span::def_id::DefIdSet; use rustc_span::{SourceFile, StableSourceFileId}; use tracing::debug; @@ -24,6 +20,7 @@ use crate::llvm; mod covfun; mod spans; +mod unused; /// Generates and exports the coverage map, which is embedded in special /// linker sections in the final binary. @@ -76,12 +73,11 @@ pub(crate) fn finalize(cx: &CodegenCx<'_, '_>) { // In a single designated CGU, also prepare covfun records for functions // in this crate that were instrumented for coverage, but are unused. if cx.codegen_unit.is_code_coverage_dead_code_cgu() { - let mut unused_instances = gather_unused_function_instances(cx); - // Sort the unused instances by symbol name, for the same reason as the used ones. - unused_instances.sort_by_cached_key(|&instance| tcx.symbol_name(instance).name); - covfun_records.extend(unused_instances.into_iter().filter_map(|instance| { - prepare_covfun_record(tcx, &mut global_file_table, instance, false) - })); + unused::prepare_covfun_records_for_unused_functions( + cx, + &mut global_file_table, + &mut covfun_records, + ); } // If there are no covfun records for this CGU, don't generate a covmap record. @@ -100,33 +96,10 @@ pub(crate) fn finalize(cx: &CodegenCx<'_, '_>) { // contain multiple covmap records from different compilation units. let filenames_hash = llvm_cov::hash_bytes(&filenames_buffer); - let mut unused_function_names = vec![]; - for covfun in &covfun_records { - unused_function_names.extend(covfun.mangled_function_name_if_unused()); - covfun::generate_covfun_record(cx, filenames_hash, covfun) } - // For unused functions, we need to take their mangled names and store them - // in a specially-named global array. LLVM's `InstrProfiling` pass will - // detect this global and include those names in its `__llvm_prf_names` - // section. (See `llvm/lib/Transforms/Instrumentation/InstrProfiling.cpp`.) - if !unused_function_names.is_empty() { - assert!(cx.codegen_unit.is_code_coverage_dead_code_cgu()); - - let name_globals = unused_function_names - .into_iter() - .map(|mangled_function_name| cx.const_str(mangled_function_name).0) - .collect::<Vec<_>>(); - let initializer = cx.const_array(cx.type_ptr(), &name_globals); - - let array = llvm::add_global(cx.llmod, cx.val_ty(initializer), c"__llvm_coverage_names"); - llvm::set_global_constant(array, true); - llvm::set_linkage(array, llvm::Linkage::InternalLinkage); - llvm::set_initializer(array, initializer); - } - // Generate the coverage map header, which contains the filenames used by // this CGU's coverage mappings, and store it in a well-known global. // (This is skipped if we returned early due to having no covfun records.) @@ -249,121 +222,3 @@ fn generate_covmap_record<'ll>(cx: &CodegenCx<'ll, '_>, version: u32, filenames_ cx.add_used_global(covmap_global); } - -/// Each CGU will normally only emit coverage metadata for the functions that it actually generates. -/// But since we don't want unused functions to disappear from coverage reports, we also scan for -/// functions that were instrumented but are not participating in codegen. -/// -/// These unused functions don't need to be codegenned, but we do need to add them to the function -/// coverage map (in a single designated CGU) so that we still emit coverage mappings for them. -/// We also end up adding their symbol names to a special global array that LLVM will include in -/// its embedded coverage data. -fn gather_unused_function_instances<'tcx>(cx: &CodegenCx<'_, 'tcx>) -> Vec<ty::Instance<'tcx>> { - assert!(cx.codegen_unit.is_code_coverage_dead_code_cgu()); - - let tcx = cx.tcx; - let usage = prepare_usage_sets(tcx); - - let is_unused_fn = |def_id: LocalDefId| -> bool { - // Usage sets expect `DefId`, so convert from `LocalDefId`. - let d: DefId = LocalDefId::to_def_id(def_id); - // To be potentially eligible for "unused function" mappings, a definition must: - // - Be eligible for coverage instrumentation - // - Not participate directly in codegen (or have lost all its coverage statements) - // - Not have any coverage statements inlined into codegenned functions - tcx.is_eligible_for_coverage(def_id) - && (!usage.all_mono_items.contains(&d) || usage.missing_own_coverage.contains(&d)) - && !usage.used_via_inlining.contains(&d) - }; - - // FIXME(#79651): Consider trying to filter out dummy instantiations of - // unused generic functions from library crates, because they can produce - // "unused instantiation" in coverage reports even when they are actually - // used by some downstream crate in the same binary. - - tcx.mir_keys(()) - .iter() - .copied() - .filter(|&def_id| is_unused_fn(def_id)) - .map(|def_id| make_dummy_instance(tcx, def_id)) - .collect::<Vec<_>>() -} - -struct UsageSets<'tcx> { - all_mono_items: &'tcx DefIdSet, - used_via_inlining: FxHashSet<DefId>, - missing_own_coverage: FxHashSet<DefId>, -} - -/// Prepare sets of definitions that are relevant to deciding whether something -/// is an "unused function" for coverage purposes. -fn prepare_usage_sets<'tcx>(tcx: TyCtxt<'tcx>) -> UsageSets<'tcx> { - let MonoItemPartitions { all_mono_items, codegen_units, .. } = - tcx.collect_and_partition_mono_items(()); - - // Obtain a MIR body for each function participating in codegen, via an - // arbitrary instance. - let mut def_ids_seen = FxHashSet::default(); - let def_and_mir_for_all_mono_fns = codegen_units - .iter() - .flat_map(|cgu| cgu.items().keys()) - .filter_map(|item| match item { - mir::mono::MonoItem::Fn(instance) => Some(instance), - mir::mono::MonoItem::Static(_) | mir::mono::MonoItem::GlobalAsm(_) => None, - }) - // We only need one arbitrary instance per definition. - .filter(move |instance| def_ids_seen.insert(instance.def_id())) - .map(|instance| { - // We don't care about the instance, just its underlying MIR. - let body = tcx.instance_mir(instance.def); - (instance.def_id(), body) - }); - - // Functions whose coverage statements were found inlined into other functions. - let mut used_via_inlining = FxHashSet::default(); - // Functions that were instrumented, but had all of their coverage statements - // removed by later MIR transforms (e.g. UnreachablePropagation). - let mut missing_own_coverage = FxHashSet::default(); - - for (def_id, body) in def_and_mir_for_all_mono_fns { - let mut saw_own_coverage = false; - - // Inspect every coverage statement in the function's MIR. - for stmt in body - .basic_blocks - .iter() - .flat_map(|block| &block.statements) - .filter(|stmt| matches!(stmt.kind, mir::StatementKind::Coverage(_))) - { - if let Some(inlined) = stmt.source_info.scope.inlined_instance(&body.source_scopes) { - // This coverage statement was inlined from another function. - used_via_inlining.insert(inlined.def_id()); - } else { - // Non-inlined coverage statements belong to the enclosing function. - saw_own_coverage = true; - } - } - - if !saw_own_coverage && body.function_coverage_info.is_some() { - missing_own_coverage.insert(def_id); - } - } - - UsageSets { all_mono_items, used_via_inlining, missing_own_coverage } -} - -fn make_dummy_instance<'tcx>(tcx: TyCtxt<'tcx>, local_def_id: LocalDefId) -> ty::Instance<'tcx> { - let def_id = local_def_id.to_def_id(); - - // Make a dummy instance that fills in all generics with placeholders. - ty::Instance::new( - def_id, - ty::GenericArgs::for_item(tcx, def_id, |param, _| { - if let ty::GenericParamDefKind::Lifetime = param.kind { - tcx.lifetimes.re_erased.into() - } else { - tcx.mk_param_from_def(param) - } - }), - ) -} diff --git a/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen/covfun.rs b/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen/covfun.rs index 048e1988c32..93419c2caad 100644 --- a/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen/covfun.rs +++ b/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen/covfun.rs @@ -37,14 +37,6 @@ pub(crate) struct CovfunRecord<'tcx> { regions: ffi::Regions, } -impl<'tcx> CovfunRecord<'tcx> { - /// FIXME(Zalathar): Make this the responsibility of the code that determines - /// which functions are unused. - pub(crate) fn mangled_function_name_if_unused(&self) -> Option<&'tcx str> { - (!self.is_used).then_some(self.mangled_function_name) - } -} - pub(crate) fn prepare_covfun_record<'tcx>( tcx: TyCtxt<'tcx>, global_file_table: &mut GlobalFileTable, diff --git a/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen/unused.rs b/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen/unused.rs new file mode 100644 index 00000000000..ab030f5b615 --- /dev/null +++ b/compiler/rustc_codegen_llvm/src/coverageinfo/mapgen/unused.rs @@ -0,0 +1,172 @@ +use rustc_codegen_ssa::traits::{BaseTypeCodegenMethods, ConstCodegenMethods}; +use rustc_data_structures::fx::FxHashSet; +use rustc_hir::def_id::{DefId, LocalDefId}; +use rustc_middle::mir; +use rustc_middle::mir::mono::MonoItemPartitions; +use rustc_middle::ty::{self, TyCtxt}; +use rustc_span::def_id::DefIdSet; + +use crate::common::CodegenCx; +use crate::coverageinfo::mapgen::GlobalFileTable; +use crate::coverageinfo::mapgen::covfun::{CovfunRecord, prepare_covfun_record}; +use crate::llvm; + +/// Each CGU will normally only emit coverage metadata for the functions that it actually generates. +/// But since we don't want unused functions to disappear from coverage reports, we also scan for +/// functions that were instrumented but are not participating in codegen. +/// +/// These unused functions don't need to be codegenned, but we do need to add them to the function +/// coverage map (in a single designated CGU) so that we still emit coverage mappings for them. +/// We also end up adding their symbol names to a special global array that LLVM will include in +/// its embedded coverage data. +pub(crate) fn prepare_covfun_records_for_unused_functions<'tcx>( + cx: &CodegenCx<'_, 'tcx>, + global_file_table: &mut GlobalFileTable, + covfun_records: &mut Vec<CovfunRecord<'tcx>>, +) { + assert!(cx.codegen_unit.is_code_coverage_dead_code_cgu()); + + let mut unused_instances = gather_unused_function_instances(cx); + // Sort the unused instances by symbol name, so that their order isn't hash-sensitive. + unused_instances.sort_by_key(|instance| instance.symbol_name); + + // Try to create a covfun record for each unused function. + let mut name_globals = Vec::with_capacity(unused_instances.len()); + covfun_records.extend(unused_instances.into_iter().filter_map(|unused| try { + let record = prepare_covfun_record(cx.tcx, global_file_table, unused.instance, false)?; + // If successful, also store its symbol name in a global constant. + name_globals.push(cx.const_str(unused.symbol_name.name).0); + record + })); + + // Store the names of unused functions in a specially-named global array. + // LLVM's `InstrProfilling` pass will detect this array, and include the + // referenced names in its `__llvm_prf_names` section. + // (See `llvm/lib/Transforms/Instrumentation/InstrProfiling.cpp`.) + if !name_globals.is_empty() { + let initializer = cx.const_array(cx.type_ptr(), &name_globals); + + let array = llvm::add_global(cx.llmod, cx.val_ty(initializer), c"__llvm_coverage_names"); + llvm::set_global_constant(array, true); + llvm::set_linkage(array, llvm::Linkage::InternalLinkage); + llvm::set_initializer(array, initializer); + } +} + +/// Holds a dummy function instance along with its symbol name, to avoid having +/// to repeatedly query for the name. +struct UnusedInstance<'tcx> { + instance: ty::Instance<'tcx>, + symbol_name: ty::SymbolName<'tcx>, +} + +fn gather_unused_function_instances<'tcx>(cx: &CodegenCx<'_, 'tcx>) -> Vec<UnusedInstance<'tcx>> { + assert!(cx.codegen_unit.is_code_coverage_dead_code_cgu()); + + let tcx = cx.tcx; + let usage = prepare_usage_sets(tcx); + + let is_unused_fn = |def_id: LocalDefId| -> bool { + // Usage sets expect `DefId`, so convert from `LocalDefId`. + let d: DefId = LocalDefId::to_def_id(def_id); + // To be potentially eligible for "unused function" mappings, a definition must: + // - Be eligible for coverage instrumentation + // - Not participate directly in codegen (or have lost all its coverage statements) + // - Not have any coverage statements inlined into codegenned functions + tcx.is_eligible_for_coverage(def_id) + && (!usage.all_mono_items.contains(&d) || usage.missing_own_coverage.contains(&d)) + && !usage.used_via_inlining.contains(&d) + }; + + // FIXME(#79651): Consider trying to filter out dummy instantiations of + // unused generic functions from library crates, because they can produce + // "unused instantiation" in coverage reports even when they are actually + // used by some downstream crate in the same binary. + + tcx.mir_keys(()) + .iter() + .copied() + .filter(|&def_id| is_unused_fn(def_id)) + .map(|def_id| make_dummy_instance(tcx, def_id)) + .map(|instance| UnusedInstance { instance, symbol_name: tcx.symbol_name(instance) }) + .collect::<Vec<_>>() +} + +struct UsageSets<'tcx> { + all_mono_items: &'tcx DefIdSet, + used_via_inlining: FxHashSet<DefId>, + missing_own_coverage: FxHashSet<DefId>, +} + +/// Prepare sets of definitions that are relevant to deciding whether something +/// is an "unused function" for coverage purposes. +fn prepare_usage_sets<'tcx>(tcx: TyCtxt<'tcx>) -> UsageSets<'tcx> { + let MonoItemPartitions { all_mono_items, codegen_units, .. } = + tcx.collect_and_partition_mono_items(()); + + // Obtain a MIR body for each function participating in codegen, via an + // arbitrary instance. + let mut def_ids_seen = FxHashSet::default(); + let def_and_mir_for_all_mono_fns = codegen_units + .iter() + .flat_map(|cgu| cgu.items().keys()) + .filter_map(|item| match item { + mir::mono::MonoItem::Fn(instance) => Some(instance), + mir::mono::MonoItem::Static(_) | mir::mono::MonoItem::GlobalAsm(_) => None, + }) + // We only need one arbitrary instance per definition. + .filter(move |instance| def_ids_seen.insert(instance.def_id())) + .map(|instance| { + // We don't care about the instance, just its underlying MIR. + let body = tcx.instance_mir(instance.def); + (instance.def_id(), body) + }); + + // Functions whose coverage statements were found inlined into other functions. + let mut used_via_inlining = FxHashSet::default(); + // Functions that were instrumented, but had all of their coverage statements + // removed by later MIR transforms (e.g. UnreachablePropagation). + let mut missing_own_coverage = FxHashSet::default(); + + for (def_id, body) in def_and_mir_for_all_mono_fns { + let mut saw_own_coverage = false; + + // Inspect every coverage statement in the function's MIR. + for stmt in body + .basic_blocks + .iter() + .flat_map(|block| &block.statements) + .filter(|stmt| matches!(stmt.kind, mir::StatementKind::Coverage(_))) + { + if let Some(inlined) = stmt.source_info.scope.inlined_instance(&body.source_scopes) { + // This coverage statement was inlined from another function. + used_via_inlining.insert(inlined.def_id()); + } else { + // Non-inlined coverage statements belong to the enclosing function. + saw_own_coverage = true; + } + } + + if !saw_own_coverage && body.function_coverage_info.is_some() { + missing_own_coverage.insert(def_id); + } + } + + UsageSets { all_mono_items, used_via_inlining, missing_own_coverage } +} + +fn make_dummy_instance<'tcx>(tcx: TyCtxt<'tcx>, local_def_id: LocalDefId) -> ty::Instance<'tcx> { + let def_id = local_def_id.to_def_id(); + + // Make a dummy instance that fills in all generics with placeholders. + ty::Instance::new( + def_id, + ty::GenericArgs::for_item(tcx, def_id, |param, _| { + if let ty::GenericParamDefKind::Lifetime = param.kind { + tcx.lifetimes.re_erased.into() + } else { + tcx.mk_param_from_def(param) + } + }), + ) +} diff --git a/compiler/rustc_hir/src/hir.rs b/compiler/rustc_hir/src/hir.rs index 1a6c15b66a4..74957b4651a 100644 --- a/compiler/rustc_hir/src/hir.rs +++ b/compiler/rustc_hir/src/hir.rs @@ -1555,6 +1555,7 @@ impl<'hir> Pat<'hir> { use PatKind::*; match self.kind { + Missing => unreachable!(), Wild | Never | Expr(_) | Range(..) | Binding(.., None) | Err(_) => true, Box(s) | Deref(s) | Ref(s, _) | Binding(.., Some(s)) | Guard(s, _) => s.walk_short_(it), Struct(_, fields, _) => fields.iter().all(|field| field.pat.walk_short_(it)), @@ -1582,7 +1583,7 @@ impl<'hir> Pat<'hir> { use PatKind::*; match self.kind { - Wild | Never | Expr(_) | Range(..) | Binding(.., None) | Err(_) => {} + Missing | Wild | Never | Expr(_) | Range(..) | Binding(.., None) | Err(_) => {} Box(s) | Deref(s) | Ref(s, _) | Binding(.., Some(s)) | Guard(s, _) => s.walk_(it), Struct(_, fields, _) => fields.iter().for_each(|field| field.pat.walk_(it)), TupleStruct(_, s, _) | Tuple(s, _) | Or(s) => s.iter().for_each(|p| p.walk_(it)), @@ -1720,6 +1721,9 @@ pub enum TyPatKind<'hir> { #[derive(Debug, Clone, Copy, HashStable_Generic)] pub enum PatKind<'hir> { + /// A missing pattern, e.g. for an anonymous param in a bare fn like `fn f(u32)`. + Missing, + /// Represents a wildcard pattern (i.e., `_`). Wild, diff --git a/compiler/rustc_hir/src/intravisit.rs b/compiler/rustc_hir/src/intravisit.rs index 506358341b5..ea3f396761b 100644 --- a/compiler/rustc_hir/src/intravisit.rs +++ b/compiler/rustc_hir/src/intravisit.rs @@ -744,7 +744,7 @@ pub fn walk_pat<'v, V: Visitor<'v>>(visitor: &mut V, pattern: &'v Pat<'v>) -> V: visit_opt!(visitor, visit_pat_expr, lower_bound); visit_opt!(visitor, visit_pat_expr, upper_bound); } - PatKind::Never | PatKind::Wild | PatKind::Err(_) => (), + PatKind::Missing | PatKind::Never | PatKind::Wild | PatKind::Err(_) => (), PatKind::Slice(prepatterns, ref slice_pattern, postpatterns) => { walk_list!(visitor, visit_pat, prepatterns); visit_opt!(visitor, visit_pat, slice_pattern); diff --git a/compiler/rustc_hir_analysis/src/check/region.rs b/compiler/rustc_hir_analysis/src/check/region.rs index cf66ab708bb..d230f9f0fc6 100644 --- a/compiler/rustc_hir_analysis/src/check/region.rs +++ b/compiler/rustc_hir_analysis/src/check/region.rs @@ -446,14 +446,11 @@ fn resolve_expr<'tcx>( // Mark this expr's scope and all parent scopes as containing `yield`. let mut scope = Scope { local_id: expr.hir_id.local_id, data: ScopeData::Node }; loop { - let span = match expr.kind { - hir::ExprKind::Yield(expr, hir::YieldSource::Await { .. }) => { - expr.span.shrink_to_hi().to(expr.span) - } - _ => expr.span, + let data = YieldData { + span: expr.span, + expr_and_pat_count: visitor.expr_and_pat_count, + source: *source, }; - let data = - YieldData { span, expr_and_pat_count: visitor.expr_and_pat_count, source: *source }; match visitor.scope_tree.yield_in_scope.get_mut(&scope) { Some(yields) => yields.push(data), None => { @@ -626,6 +623,7 @@ fn resolve_local<'tcx>( PatKind::Ref(_, _) | PatKind::Binding(hir::BindingMode(hir::ByRef::No, _), ..) + | PatKind::Missing | PatKind::Wild | PatKind::Never | PatKind::Expr(_) diff --git a/compiler/rustc_hir_pretty/src/lib.rs b/compiler/rustc_hir_pretty/src/lib.rs index 8c0c17f7a7d..29214196c3d 100644 --- a/compiler/rustc_hir_pretty/src/lib.rs +++ b/compiler/rustc_hir_pretty/src/lib.rs @@ -1868,6 +1868,7 @@ impl<'a> State<'a> { // Pat isn't normalized, but the beauty of it // is that it doesn't matter match pat.kind { + PatKind::Missing => unreachable!(), PatKind::Wild => self.word("_"), PatKind::Never => self.word("!"), PatKind::Binding(BindingMode(by_ref, mutbl), _, ident, sub) => { diff --git a/compiler/rustc_hir_typeck/src/demand.rs b/compiler/rustc_hir_typeck/src/demand.rs index b845e2190ef..fec45995410 100644 --- a/compiler/rustc_hir_typeck/src/demand.rs +++ b/compiler/rustc_hir_typeck/src/demand.rs @@ -865,10 +865,7 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> { // `&'name Ty` -> `&'name mut Ty` or `&Ty` -> `&mut Ty` vec![( ty_ref.1.ty.span.shrink_to_lo(), - format!( - "{}mut ", - if ty_ref.0.ident.span.lo() == ty_ref.0.ident.span.hi() { "" } else { " " }, - ), + format!("{}mut ", if ty_ref.0.ident.span.is_empty() { "" } else { " " },), )] }; sugg.extend([ diff --git a/compiler/rustc_hir_typeck/src/expr.rs b/compiler/rustc_hir_typeck/src/expr.rs index 45ab8e03db5..952a2e231e4 100644 --- a/compiler/rustc_hir_typeck/src/expr.rs +++ b/compiler/rustc_hir_typeck/src/expr.rs @@ -482,7 +482,8 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> { // All of these constitute a read, or match on something that isn't `!`, // which would require a `NeverToAny` coercion. - hir::PatKind::Binding(_, _, _, _) + hir::PatKind::Missing + | hir::PatKind::Binding(_, _, _, _) | hir::PatKind::Struct(_, _, _) | hir::PatKind::TupleStruct(_, _, _) | hir::PatKind::Tuple(_, _) diff --git a/compiler/rustc_hir_typeck/src/expr_use_visitor.rs b/compiler/rustc_hir_typeck/src/expr_use_visitor.rs index 64176dacb73..27df8f0e98a 100644 --- a/compiler/rustc_hir_typeck/src/expr_use_visitor.rs +++ b/compiler/rustc_hir_typeck/src/expr_use_visitor.rs @@ -611,6 +611,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx for pat in pats { self.cat_pattern(discr_place.clone(), pat, &mut |place, pat| { match &pat.kind { + PatKind::Missing => unreachable!(), PatKind::Binding(.., opt_sub_pat) => { // If the opt_sub_pat is None, then the binding does not count as // a wildcard for the purpose of borrowing discr. @@ -1832,6 +1833,7 @@ impl<'tcx, Cx: TypeInformationCtxt<'tcx>, D: Delegate<'tcx>> ExprUseVisitor<'tcx | PatKind::Expr(..) | PatKind::Range(..) | PatKind::Never + | PatKind::Missing | PatKind::Wild | PatKind::Err(_) => { // always ok diff --git a/compiler/rustc_hir_typeck/src/pat.rs b/compiler/rustc_hir_typeck/src/pat.rs index 9766ceda569..8641348bffb 100644 --- a/compiler/rustc_hir_typeck/src/pat.rs +++ b/compiler/rustc_hir_typeck/src/pat.rs @@ -341,7 +341,7 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> { }; let ty = match pat.kind { - PatKind::Wild | PatKind::Err(_) => expected, + PatKind::Missing | PatKind::Wild | PatKind::Err(_) => expected, // We allow any type here; we ensure that the type is uninhabited during match checking. PatKind::Never => expected, PatKind::Expr(PatExpr { kind: PatExprKind::Path(qpath), hir_id, span }) => { @@ -505,9 +505,11 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> { }, // Ref patterns are complicated, we handle them in `check_pat_ref`. - PatKind::Ref(..) => AdjustMode::Pass, + PatKind::Ref(..) + // No need to do anything on a missing pattern. + | PatKind::Missing // A `_` pattern works with any expected type, so there's no need to do anything. - PatKind::Wild + | PatKind::Wild // A malformed pattern doesn't have an expected type, so let's just accept any type. | PatKind::Err(_) // Bindings also work with whatever the expected type is, @@ -1032,7 +1034,8 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> { | PatKind::Tuple(..) | PatKind::Slice(..) => "binding", - PatKind::Wild + PatKind::Missing + | PatKind::Wild | PatKind::Never | PatKind::Binding(..) | PatKind::Box(..) diff --git a/compiler/rustc_lexer/src/lib.rs b/compiler/rustc_lexer/src/lib.rs index 61638e45253..f9c71b2fa65 100644 --- a/compiler/rustc_lexer/src/lib.rs +++ b/compiler/rustc_lexer/src/lib.rs @@ -26,7 +26,6 @@ // tidy-alphabetical-end mod cursor; -pub mod unescape; #[cfg(test)] mod tests; diff --git a/compiler/rustc_lexer/src/unescape.rs b/compiler/rustc_lexer/src/unescape.rs deleted file mode 100644 index d6ea4249247..00000000000 --- a/compiler/rustc_lexer/src/unescape.rs +++ /dev/null @@ -1,438 +0,0 @@ -//! Utilities for validating string and char literals and turning them into -//! values they represent. - -use std::ops::Range; -use std::str::Chars; - -use Mode::*; - -#[cfg(test)] -mod tests; - -/// Errors and warnings that can occur during string unescaping. They mostly -/// relate to malformed escape sequences, but there are a few that are about -/// other problems. -#[derive(Debug, PartialEq, Eq)] -pub enum EscapeError { - /// Expected 1 char, but 0 were found. - ZeroChars, - /// Expected 1 char, but more than 1 were found. - MoreThanOneChar, - - /// Escaped '\' character without continuation. - LoneSlash, - /// Invalid escape character (e.g. '\z'). - InvalidEscape, - /// Raw '\r' encountered. - BareCarriageReturn, - /// Raw '\r' encountered in raw string. - BareCarriageReturnInRawString, - /// Unescaped character that was expected to be escaped (e.g. raw '\t'). - EscapeOnlyChar, - - /// Numeric character escape is too short (e.g. '\x1'). - TooShortHexEscape, - /// Invalid character in numeric escape (e.g. '\xz') - InvalidCharInHexEscape, - /// Character code in numeric escape is non-ascii (e.g. '\xFF'). - OutOfRangeHexEscape, - - /// '\u' not followed by '{'. - NoBraceInUnicodeEscape, - /// Non-hexadecimal value in '\u{..}'. - InvalidCharInUnicodeEscape, - /// '\u{}' - EmptyUnicodeEscape, - /// No closing brace in '\u{..}', e.g. '\u{12'. - UnclosedUnicodeEscape, - /// '\u{_12}' - LeadingUnderscoreUnicodeEscape, - /// More than 6 characters in '\u{..}', e.g. '\u{10FFFF_FF}' - OverlongUnicodeEscape, - /// Invalid in-bound unicode character code, e.g. '\u{DFFF}'. - LoneSurrogateUnicodeEscape, - /// Out of bounds unicode character code, e.g. '\u{FFFFFF}'. - OutOfRangeUnicodeEscape, - - /// Unicode escape code in byte literal. - UnicodeEscapeInByte, - /// Non-ascii character in byte literal, byte string literal, or raw byte string literal. - NonAsciiCharInByte, - - // `\0` in a C string literal. - NulInCStr, - - /// After a line ending with '\', the next line contains whitespace - /// characters that are not skipped. - UnskippedWhitespaceWarning, - - /// After a line ending with '\', multiple lines are skipped. - MultipleSkippedLinesWarning, -} - -impl EscapeError { - /// Returns true for actual errors, as opposed to warnings. - pub fn is_fatal(&self) -> bool { - !matches!( - self, - EscapeError::UnskippedWhitespaceWarning | EscapeError::MultipleSkippedLinesWarning - ) - } -} - -/// Takes the contents of a unicode-only (non-mixed-utf8) literal (without -/// quotes) and produces a sequence of escaped characters or errors. -/// -/// Values are returned by invoking `callback`. For `Char` and `Byte` modes, -/// the callback will be called exactly once. -pub fn unescape_unicode<F>(src: &str, mode: Mode, callback: &mut F) -where - F: FnMut(Range<usize>, Result<char, EscapeError>), -{ - match mode { - Char | Byte => { - let mut chars = src.chars(); - let res = unescape_char_or_byte(&mut chars, mode); - callback(0..(src.len() - chars.as_str().len()), res); - } - Str | ByteStr => unescape_non_raw_common(src, mode, callback), - RawStr | RawByteStr => check_raw_common(src, mode, callback), - RawCStr => check_raw_common(src, mode, &mut |r, mut result| { - if let Ok('\0') = result { - result = Err(EscapeError::NulInCStr); - } - callback(r, result) - }), - CStr => unreachable!(), - } -} - -/// Used for mixed utf8 string literals, i.e. those that allow both unicode -/// chars and high bytes. -pub enum MixedUnit { - /// Used for ASCII chars (written directly or via `\x00`..`\x7f` escapes) - /// and Unicode chars (written directly or via `\u` escapes). - /// - /// For example, if '¥' appears in a string it is represented here as - /// `MixedUnit::Char('¥')`, and it will be appended to the relevant byte - /// string as the two-byte UTF-8 sequence `[0xc2, 0xa5]` - Char(char), - - /// Used for high bytes (`\x80`..`\xff`). - /// - /// For example, if `\xa5` appears in a string it is represented here as - /// `MixedUnit::HighByte(0xa5)`, and it will be appended to the relevant - /// byte string as the single byte `0xa5`. - HighByte(u8), -} - -impl From<char> for MixedUnit { - fn from(c: char) -> Self { - MixedUnit::Char(c) - } -} - -impl From<u8> for MixedUnit { - fn from(n: u8) -> Self { - if n.is_ascii() { MixedUnit::Char(n as char) } else { MixedUnit::HighByte(n) } - } -} - -/// Takes the contents of a mixed-utf8 literal (without quotes) and produces -/// a sequence of escaped characters or errors. -/// -/// Values are returned by invoking `callback`. -pub fn unescape_mixed<F>(src: &str, mode: Mode, callback: &mut F) -where - F: FnMut(Range<usize>, Result<MixedUnit, EscapeError>), -{ - match mode { - CStr => unescape_non_raw_common(src, mode, &mut |r, mut result| { - if let Ok(MixedUnit::Char('\0')) = result { - result = Err(EscapeError::NulInCStr); - } - callback(r, result) - }), - Char | Byte | Str | RawStr | ByteStr | RawByteStr | RawCStr => unreachable!(), - } -} - -/// Takes a contents of a char literal (without quotes), and returns an -/// unescaped char or an error. -pub fn unescape_char(src: &str) -> Result<char, EscapeError> { - unescape_char_or_byte(&mut src.chars(), Char) -} - -/// Takes a contents of a byte literal (without quotes), and returns an -/// unescaped byte or an error. -pub fn unescape_byte(src: &str) -> Result<u8, EscapeError> { - unescape_char_or_byte(&mut src.chars(), Byte).map(byte_from_char) -} - -/// What kind of literal do we parse. -#[derive(Debug, Clone, Copy, PartialEq)] -pub enum Mode { - Char, - - Byte, - - Str, - RawStr, - - ByteStr, - RawByteStr, - - CStr, - RawCStr, -} - -impl Mode { - pub fn in_double_quotes(self) -> bool { - match self { - Str | RawStr | ByteStr | RawByteStr | CStr | RawCStr => true, - Char | Byte => false, - } - } - - /// Are `\x80`..`\xff` allowed? - fn allow_high_bytes(self) -> bool { - match self { - Char | Str => false, - Byte | ByteStr | CStr => true, - RawStr | RawByteStr | RawCStr => unreachable!(), - } - } - - /// Are unicode (non-ASCII) chars allowed? - #[inline] - fn allow_unicode_chars(self) -> bool { - match self { - Byte | ByteStr | RawByteStr => false, - Char | Str | RawStr | CStr | RawCStr => true, - } - } - - /// Are unicode escapes (`\u`) allowed? - fn allow_unicode_escapes(self) -> bool { - match self { - Byte | ByteStr => false, - Char | Str | CStr => true, - RawByteStr | RawStr | RawCStr => unreachable!(), - } - } - - pub fn prefix_noraw(self) -> &'static str { - match self { - Char | Str | RawStr => "", - Byte | ByteStr | RawByteStr => "b", - CStr | RawCStr => "c", - } - } -} - -fn scan_escape<T: From<char> + From<u8>>( - chars: &mut Chars<'_>, - mode: Mode, -) -> Result<T, EscapeError> { - // Previous character was '\\', unescape what follows. - let res: char = match chars.next().ok_or(EscapeError::LoneSlash)? { - '"' => '"', - 'n' => '\n', - 'r' => '\r', - 't' => '\t', - '\\' => '\\', - '\'' => '\'', - '0' => '\0', - 'x' => { - // Parse hexadecimal character code. - - let hi = chars.next().ok_or(EscapeError::TooShortHexEscape)?; - let hi = hi.to_digit(16).ok_or(EscapeError::InvalidCharInHexEscape)?; - - let lo = chars.next().ok_or(EscapeError::TooShortHexEscape)?; - let lo = lo.to_digit(16).ok_or(EscapeError::InvalidCharInHexEscape)?; - - let value = (hi * 16 + lo) as u8; - - return if !mode.allow_high_bytes() && !value.is_ascii() { - Err(EscapeError::OutOfRangeHexEscape) - } else { - // This may be a high byte, but that will only happen if `T` is - // `MixedUnit`, because of the `allow_high_bytes` check above. - Ok(T::from(value)) - }; - } - 'u' => return scan_unicode(chars, mode.allow_unicode_escapes()).map(T::from), - _ => return Err(EscapeError::InvalidEscape), - }; - Ok(T::from(res)) -} - -fn scan_unicode(chars: &mut Chars<'_>, allow_unicode_escapes: bool) -> Result<char, EscapeError> { - // We've parsed '\u', now we have to parse '{..}'. - - if chars.next() != Some('{') { - return Err(EscapeError::NoBraceInUnicodeEscape); - } - - // First character must be a hexadecimal digit. - let mut n_digits = 1; - let mut value: u32 = match chars.next().ok_or(EscapeError::UnclosedUnicodeEscape)? { - '_' => return Err(EscapeError::LeadingUnderscoreUnicodeEscape), - '}' => return Err(EscapeError::EmptyUnicodeEscape), - c => c.to_digit(16).ok_or(EscapeError::InvalidCharInUnicodeEscape)?, - }; - - // First character is valid, now parse the rest of the number - // and closing brace. - loop { - match chars.next() { - None => return Err(EscapeError::UnclosedUnicodeEscape), - Some('_') => continue, - Some('}') => { - if n_digits > 6 { - return Err(EscapeError::OverlongUnicodeEscape); - } - - // Incorrect syntax has higher priority for error reporting - // than unallowed value for a literal. - if !allow_unicode_escapes { - return Err(EscapeError::UnicodeEscapeInByte); - } - - break std::char::from_u32(value).ok_or({ - if value > 0x10FFFF { - EscapeError::OutOfRangeUnicodeEscape - } else { - EscapeError::LoneSurrogateUnicodeEscape - } - }); - } - Some(c) => { - let digit: u32 = c.to_digit(16).ok_or(EscapeError::InvalidCharInUnicodeEscape)?; - n_digits += 1; - if n_digits > 6 { - // Stop updating value since we're sure that it's incorrect already. - continue; - } - value = value * 16 + digit; - } - }; - } -} - -#[inline] -fn ascii_check(c: char, allow_unicode_chars: bool) -> Result<char, EscapeError> { - if allow_unicode_chars || c.is_ascii() { Ok(c) } else { Err(EscapeError::NonAsciiCharInByte) } -} - -fn unescape_char_or_byte(chars: &mut Chars<'_>, mode: Mode) -> Result<char, EscapeError> { - let c = chars.next().ok_or(EscapeError::ZeroChars)?; - let res = match c { - '\\' => scan_escape(chars, mode), - '\n' | '\t' | '\'' => Err(EscapeError::EscapeOnlyChar), - '\r' => Err(EscapeError::BareCarriageReturn), - _ => ascii_check(c, mode.allow_unicode_chars()), - }?; - if chars.next().is_some() { - return Err(EscapeError::MoreThanOneChar); - } - Ok(res) -} - -/// Takes a contents of a string literal (without quotes) and produces a -/// sequence of escaped characters or errors. -fn unescape_non_raw_common<F, T: From<char> + From<u8>>(src: &str, mode: Mode, callback: &mut F) -where - F: FnMut(Range<usize>, Result<T, EscapeError>), -{ - let mut chars = src.chars(); - let allow_unicode_chars = mode.allow_unicode_chars(); // get this outside the loop - - // The `start` and `end` computation here is complicated because - // `skip_ascii_whitespace` makes us to skip over chars without counting - // them in the range computation. - while let Some(c) = chars.next() { - let start = src.len() - chars.as_str().len() - c.len_utf8(); - let res = match c { - '\\' => { - match chars.clone().next() { - Some('\n') => { - // Rust language specification requires us to skip whitespaces - // if unescaped '\' character is followed by '\n'. - // For details see [Rust language reference] - // (https://doc.rust-lang.org/reference/tokens.html#string-literals). - skip_ascii_whitespace(&mut chars, start, &mut |range, err| { - callback(range, Err(err)) - }); - continue; - } - _ => scan_escape::<T>(&mut chars, mode), - } - } - '"' => Err(EscapeError::EscapeOnlyChar), - '\r' => Err(EscapeError::BareCarriageReturn), - _ => ascii_check(c, allow_unicode_chars).map(T::from), - }; - let end = src.len() - chars.as_str().len(); - callback(start..end, res); - } -} - -fn skip_ascii_whitespace<F>(chars: &mut Chars<'_>, start: usize, callback: &mut F) -where - F: FnMut(Range<usize>, EscapeError), -{ - let tail = chars.as_str(); - let first_non_space = tail - .bytes() - .position(|b| b != b' ' && b != b'\t' && b != b'\n' && b != b'\r') - .unwrap_or(tail.len()); - if tail[1..first_non_space].contains('\n') { - // The +1 accounts for the escaping slash. - let end = start + first_non_space + 1; - callback(start..end, EscapeError::MultipleSkippedLinesWarning); - } - let tail = &tail[first_non_space..]; - if let Some(c) = tail.chars().next() { - if c.is_whitespace() { - // For error reporting, we would like the span to contain the character that was not - // skipped. The +1 is necessary to account for the leading \ that started the escape. - let end = start + first_non_space + c.len_utf8() + 1; - callback(start..end, EscapeError::UnskippedWhitespaceWarning); - } - } - *chars = tail.chars(); -} - -/// Takes a contents of a string literal (without quotes) and produces a -/// sequence of characters or errors. -/// NOTE: Raw strings do not perform any explicit character escaping, here we -/// only produce errors on bare CR. -fn check_raw_common<F>(src: &str, mode: Mode, callback: &mut F) -where - F: FnMut(Range<usize>, Result<char, EscapeError>), -{ - let mut chars = src.chars(); - let allow_unicode_chars = mode.allow_unicode_chars(); // get this outside the loop - - // The `start` and `end` computation here matches the one in - // `unescape_non_raw_common` for consistency, even though this function - // doesn't have to worry about skipping any chars. - while let Some(c) = chars.next() { - let start = src.len() - chars.as_str().len() - c.len_utf8(); - let res = match c { - '\r' => Err(EscapeError::BareCarriageReturnInRawString), - _ => ascii_check(c, allow_unicode_chars), - }; - let end = src.len() - chars.as_str().len(); - callback(start..end, res); - } -} - -#[inline] -pub fn byte_from_char(c: char) -> u8 { - let res = c as u32; - debug_assert!(res <= u8::MAX as u32, "guaranteed because of ByteStr"); - res as u8 -} diff --git a/compiler/rustc_lexer/src/unescape/tests.rs b/compiler/rustc_lexer/src/unescape/tests.rs deleted file mode 100644 index 5b99495f475..00000000000 --- a/compiler/rustc_lexer/src/unescape/tests.rs +++ /dev/null @@ -1,286 +0,0 @@ -use super::*; - -#[test] -fn test_unescape_char_bad() { - fn check(literal_text: &str, expected_error: EscapeError) { - assert_eq!(unescape_char(literal_text), Err(expected_error)); - } - - check("", EscapeError::ZeroChars); - check(r"\", EscapeError::LoneSlash); - - check("\n", EscapeError::EscapeOnlyChar); - check("\t", EscapeError::EscapeOnlyChar); - check("'", EscapeError::EscapeOnlyChar); - check("\r", EscapeError::BareCarriageReturn); - - check("spam", EscapeError::MoreThanOneChar); - check(r"\x0ff", EscapeError::MoreThanOneChar); - check(r#"\"a"#, EscapeError::MoreThanOneChar); - check(r"\na", EscapeError::MoreThanOneChar); - check(r"\ra", EscapeError::MoreThanOneChar); - check(r"\ta", EscapeError::MoreThanOneChar); - check(r"\\a", EscapeError::MoreThanOneChar); - check(r"\'a", EscapeError::MoreThanOneChar); - check(r"\0a", EscapeError::MoreThanOneChar); - check(r"\u{0}x", EscapeError::MoreThanOneChar); - check(r"\u{1F63b}}", EscapeError::MoreThanOneChar); - - check(r"\v", EscapeError::InvalidEscape); - check(r"\💩", EscapeError::InvalidEscape); - check(r"\●", EscapeError::InvalidEscape); - check("\\\r", EscapeError::InvalidEscape); - - check(r"\x", EscapeError::TooShortHexEscape); - check(r"\x0", EscapeError::TooShortHexEscape); - check(r"\xf", EscapeError::TooShortHexEscape); - check(r"\xa", EscapeError::TooShortHexEscape); - check(r"\xx", EscapeError::InvalidCharInHexEscape); - check(r"\xы", EscapeError::InvalidCharInHexEscape); - check(r"\x🦀", EscapeError::InvalidCharInHexEscape); - check(r"\xtt", EscapeError::InvalidCharInHexEscape); - check(r"\xff", EscapeError::OutOfRangeHexEscape); - check(r"\xFF", EscapeError::OutOfRangeHexEscape); - check(r"\x80", EscapeError::OutOfRangeHexEscape); - - check(r"\u", EscapeError::NoBraceInUnicodeEscape); - check(r"\u[0123]", EscapeError::NoBraceInUnicodeEscape); - check(r"\u{0x}", EscapeError::InvalidCharInUnicodeEscape); - check(r"\u{", EscapeError::UnclosedUnicodeEscape); - check(r"\u{0000", EscapeError::UnclosedUnicodeEscape); - check(r"\u{}", EscapeError::EmptyUnicodeEscape); - check(r"\u{_0000}", EscapeError::LeadingUnderscoreUnicodeEscape); - check(r"\u{0000000}", EscapeError::OverlongUnicodeEscape); - check(r"\u{FFFFFF}", EscapeError::OutOfRangeUnicodeEscape); - check(r"\u{ffffff}", EscapeError::OutOfRangeUnicodeEscape); - check(r"\u{ffffff}", EscapeError::OutOfRangeUnicodeEscape); - - check(r"\u{DC00}", EscapeError::LoneSurrogateUnicodeEscape); - check(r"\u{DDDD}", EscapeError::LoneSurrogateUnicodeEscape); - check(r"\u{DFFF}", EscapeError::LoneSurrogateUnicodeEscape); - - check(r"\u{D800}", EscapeError::LoneSurrogateUnicodeEscape); - check(r"\u{DAAA}", EscapeError::LoneSurrogateUnicodeEscape); - check(r"\u{DBFF}", EscapeError::LoneSurrogateUnicodeEscape); -} - -#[test] -fn test_unescape_char_good() { - fn check(literal_text: &str, expected_char: char) { - assert_eq!(unescape_char(literal_text), Ok(expected_char)); - } - - check("a", 'a'); - check("ы", 'ы'); - check("🦀", '🦀'); - - check(r#"\""#, '"'); - check(r"\n", '\n'); - check(r"\r", '\r'); - check(r"\t", '\t'); - check(r"\\", '\\'); - check(r"\'", '\''); - check(r"\0", '\0'); - - check(r"\x00", '\0'); - check(r"\x5a", 'Z'); - check(r"\x5A", 'Z'); - check(r"\x7f", 127 as char); - - check(r"\u{0}", '\0'); - check(r"\u{000000}", '\0'); - check(r"\u{41}", 'A'); - check(r"\u{0041}", 'A'); - check(r"\u{00_41}", 'A'); - check(r"\u{4__1__}", 'A'); - check(r"\u{1F63b}", '😻'); -} - -#[test] -fn test_unescape_str_warn() { - fn check(literal: &str, expected: &[(Range<usize>, Result<char, EscapeError>)]) { - let mut unescaped = Vec::with_capacity(literal.len()); - unescape_unicode(literal, Mode::Str, &mut |range, res| unescaped.push((range, res))); - assert_eq!(unescaped, expected); - } - - // Check we can handle escaped newlines at the end of a file. - check("\\\n", &[]); - check("\\\n ", &[]); - - check( - "\\\n \u{a0} x", - &[ - (0..5, Err(EscapeError::UnskippedWhitespaceWarning)), - (3..5, Ok('\u{a0}')), - (5..6, Ok(' ')), - (6..7, Ok('x')), - ], - ); - check("\\\n \n x", &[(0..7, Err(EscapeError::MultipleSkippedLinesWarning)), (7..8, Ok('x'))]); -} - -#[test] -fn test_unescape_str_good() { - fn check(literal_text: &str, expected: &str) { - let mut buf = Ok(String::with_capacity(literal_text.len())); - unescape_unicode(literal_text, Mode::Str, &mut |range, c| { - if let Ok(b) = &mut buf { - match c { - Ok(c) => b.push(c), - Err(e) => buf = Err((range, e)), - } - } - }); - assert_eq!(buf.as_deref(), Ok(expected)) - } - - check("foo", "foo"); - check("", ""); - check(" \t\n", " \t\n"); - - check("hello \\\n world", "hello world"); - check("thread's", "thread's") -} - -#[test] -fn test_unescape_byte_bad() { - fn check(literal_text: &str, expected_error: EscapeError) { - assert_eq!(unescape_byte(literal_text), Err(expected_error)); - } - - check("", EscapeError::ZeroChars); - check(r"\", EscapeError::LoneSlash); - - check("\n", EscapeError::EscapeOnlyChar); - check("\t", EscapeError::EscapeOnlyChar); - check("'", EscapeError::EscapeOnlyChar); - check("\r", EscapeError::BareCarriageReturn); - - check("spam", EscapeError::MoreThanOneChar); - check(r"\x0ff", EscapeError::MoreThanOneChar); - check(r#"\"a"#, EscapeError::MoreThanOneChar); - check(r"\na", EscapeError::MoreThanOneChar); - check(r"\ra", EscapeError::MoreThanOneChar); - check(r"\ta", EscapeError::MoreThanOneChar); - check(r"\\a", EscapeError::MoreThanOneChar); - check(r"\'a", EscapeError::MoreThanOneChar); - check(r"\0a", EscapeError::MoreThanOneChar); - - check(r"\v", EscapeError::InvalidEscape); - check(r"\💩", EscapeError::InvalidEscape); - check(r"\●", EscapeError::InvalidEscape); - - check(r"\x", EscapeError::TooShortHexEscape); - check(r"\x0", EscapeError::TooShortHexEscape); - check(r"\xa", EscapeError::TooShortHexEscape); - check(r"\xf", EscapeError::TooShortHexEscape); - check(r"\xx", EscapeError::InvalidCharInHexEscape); - check(r"\xы", EscapeError::InvalidCharInHexEscape); - check(r"\x🦀", EscapeError::InvalidCharInHexEscape); - check(r"\xtt", EscapeError::InvalidCharInHexEscape); - - check(r"\u", EscapeError::NoBraceInUnicodeEscape); - check(r"\u[0123]", EscapeError::NoBraceInUnicodeEscape); - check(r"\u{0x}", EscapeError::InvalidCharInUnicodeEscape); - check(r"\u{", EscapeError::UnclosedUnicodeEscape); - check(r"\u{0000", EscapeError::UnclosedUnicodeEscape); - check(r"\u{}", EscapeError::EmptyUnicodeEscape); - check(r"\u{_0000}", EscapeError::LeadingUnderscoreUnicodeEscape); - check(r"\u{0000000}", EscapeError::OverlongUnicodeEscape); - - check("ы", EscapeError::NonAsciiCharInByte); - check("🦀", EscapeError::NonAsciiCharInByte); - - check(r"\u{0}", EscapeError::UnicodeEscapeInByte); - check(r"\u{000000}", EscapeError::UnicodeEscapeInByte); - check(r"\u{41}", EscapeError::UnicodeEscapeInByte); - check(r"\u{0041}", EscapeError::UnicodeEscapeInByte); - check(r"\u{00_41}", EscapeError::UnicodeEscapeInByte); - check(r"\u{4__1__}", EscapeError::UnicodeEscapeInByte); - check(r"\u{1F63b}", EscapeError::UnicodeEscapeInByte); - check(r"\u{0}x", EscapeError::UnicodeEscapeInByte); - check(r"\u{1F63b}}", EscapeError::UnicodeEscapeInByte); - check(r"\u{FFFFFF}", EscapeError::UnicodeEscapeInByte); - check(r"\u{ffffff}", EscapeError::UnicodeEscapeInByte); - check(r"\u{ffffff}", EscapeError::UnicodeEscapeInByte); - check(r"\u{DC00}", EscapeError::UnicodeEscapeInByte); - check(r"\u{DDDD}", EscapeError::UnicodeEscapeInByte); - check(r"\u{DFFF}", EscapeError::UnicodeEscapeInByte); - check(r"\u{D800}", EscapeError::UnicodeEscapeInByte); - check(r"\u{DAAA}", EscapeError::UnicodeEscapeInByte); - check(r"\u{DBFF}", EscapeError::UnicodeEscapeInByte); -} - -#[test] -fn test_unescape_byte_good() { - fn check(literal_text: &str, expected_byte: u8) { - assert_eq!(unescape_byte(literal_text), Ok(expected_byte)); - } - - check("a", b'a'); - - check(r#"\""#, b'"'); - check(r"\n", b'\n'); - check(r"\r", b'\r'); - check(r"\t", b'\t'); - check(r"\\", b'\\'); - check(r"\'", b'\''); - check(r"\0", b'\0'); - - check(r"\x00", b'\0'); - check(r"\x5a", b'Z'); - check(r"\x5A", b'Z'); - check(r"\x7f", 127); - check(r"\x80", 128); - check(r"\xff", 255); - check(r"\xFF", 255); -} - -#[test] -fn test_unescape_byte_str_good() { - fn check(literal_text: &str, expected: &[u8]) { - let mut buf = Ok(Vec::with_capacity(literal_text.len())); - unescape_unicode(literal_text, Mode::ByteStr, &mut |range, c| { - if let Ok(b) = &mut buf { - match c { - Ok(c) => b.push(byte_from_char(c)), - Err(e) => buf = Err((range, e)), - } - } - }); - assert_eq!(buf.as_deref(), Ok(expected)) - } - - check("foo", b"foo"); - check("", b""); - check(" \t\n", b" \t\n"); - - check("hello \\\n world", b"hello world"); - check("thread's", b"thread's") -} - -#[test] -fn test_unescape_raw_str() { - fn check(literal: &str, expected: &[(Range<usize>, Result<char, EscapeError>)]) { - let mut unescaped = Vec::with_capacity(literal.len()); - unescape_unicode(literal, Mode::RawStr, &mut |range, res| unescaped.push((range, res))); - assert_eq!(unescaped, expected); - } - - check("\r", &[(0..1, Err(EscapeError::BareCarriageReturnInRawString))]); - check("\rx", &[(0..1, Err(EscapeError::BareCarriageReturnInRawString)), (1..2, Ok('x'))]); -} - -#[test] -fn test_unescape_raw_byte_str() { - fn check(literal: &str, expected: &[(Range<usize>, Result<char, EscapeError>)]) { - let mut unescaped = Vec::with_capacity(literal.len()); - unescape_unicode(literal, Mode::RawByteStr, &mut |range, res| unescaped.push((range, res))); - assert_eq!(unescaped, expected); - } - - check("\r", &[(0..1, Err(EscapeError::BareCarriageReturnInRawString))]); - check("🦀", &[(0..4, Err(EscapeError::NonAsciiCharInByte))]); - check("🦀a", &[(0..4, Err(EscapeError::NonAsciiCharInByte)), (4..5, Ok('a'))]); -} diff --git a/compiler/rustc_lint/src/builtin.rs b/compiler/rustc_lint/src/builtin.rs index dae0efcbbc4..38c2bf54432 100644 --- a/compiler/rustc_lint/src/builtin.rs +++ b/compiler/rustc_lint/src/builtin.rs @@ -779,21 +779,19 @@ impl EarlyLintPass for AnonymousParameters { } if let ast::AssocItemKind::Fn(box Fn { ref sig, .. }) = it.kind { for arg in sig.decl.inputs.iter() { - if let ast::PatKind::Ident(_, ident, None) = arg.pat.kind { - if ident.name == kw::Empty { - let ty_snip = cx.sess().source_map().span_to_snippet(arg.ty.span); + if let ast::PatKind::Missing = arg.pat.kind { + let ty_snip = cx.sess().source_map().span_to_snippet(arg.ty.span); - let (ty_snip, appl) = if let Ok(ref snip) = ty_snip { - (snip.as_str(), Applicability::MachineApplicable) - } else { - ("<type>", Applicability::HasPlaceholders) - }; - cx.emit_span_lint( - ANONYMOUS_PARAMETERS, - arg.pat.span, - BuiltinAnonymousParams { suggestion: (arg.pat.span, appl), ty_snip }, - ); - } + let (ty_snip, appl) = if let Ok(ref snip) = ty_snip { + (snip.as_str(), Applicability::MachineApplicable) + } else { + ("<type>", Applicability::HasPlaceholders) + }; + cx.emit_span_lint( + ANONYMOUS_PARAMETERS, + arg.pat.span, + BuiltinAnonymousParams { suggestion: (arg.pat.span, appl), ty_snip }, + ); } } } diff --git a/compiler/rustc_lint/src/unused.rs b/compiler/rustc_lint/src/unused.rs index 7b43aac90c7..806bca78f78 100644 --- a/compiler/rustc_lint/src/unused.rs +++ b/compiler/rustc_lint/src/unused.rs @@ -1201,7 +1201,8 @@ impl EarlyLintPass for UnusedParens { // Do not lint on `(..)` as that will result in the other arms being useless. Paren(_) // The other cases do not contain sub-patterns. - | Wild | Never | Rest | Expr(..) | MacCall(..) | Range(..) | Ident(.., None) | Path(..) | Err(_) => {}, + | Missing | Wild | Never | Rest | Expr(..) | MacCall(..) | Range(..) | Ident(.., None) + | Path(..) | Err(_) => {}, // These are list-like patterns; parens can always be removed. TupleStruct(_, _, ps) | Tuple(ps) | Slice(ps) | Or(ps) => for p in ps { self.check_unused_parens_pat(cx, p, false, false, keep_space); diff --git a/compiler/rustc_middle/src/thir.rs b/compiler/rustc_middle/src/thir.rs index bc49c197ef8..c168142fb1e 100644 --- a/compiler/rustc_middle/src/thir.rs +++ b/compiler/rustc_middle/src/thir.rs @@ -747,6 +747,9 @@ pub struct Ascription<'tcx> { #[derive(Clone, Debug, HashStable, TypeVisitable)] pub enum PatKind<'tcx> { + /// A missing pattern, e.g. for an anonymous param in a bare fn like `fn f(u32)`. + Missing, + /// A wildcard pattern: `_`. Wild, diff --git a/compiler/rustc_middle/src/thir/visit.rs b/compiler/rustc_middle/src/thir/visit.rs index 7d62ab7970d..f3da2a5cc8e 100644 --- a/compiler/rustc_middle/src/thir/visit.rs +++ b/compiler/rustc_middle/src/thir/visit.rs @@ -250,7 +250,8 @@ pub(crate) fn for_each_immediate_subpat<'a, 'tcx>( mut callback: impl FnMut(&'a Pat<'tcx>), ) { match &pat.kind { - PatKind::Wild + PatKind::Missing + | PatKind::Wild | PatKind::Binding { subpattern: None, .. } | PatKind::Constant { value: _ } | PatKind::Range(_) diff --git a/compiler/rustc_middle/src/ty/generic_args.rs b/compiler/rustc_middle/src/ty/generic_args.rs index e87859a55ed..9c1ff134f0f 100644 --- a/compiler/rustc_middle/src/ty/generic_args.rs +++ b/compiler/rustc_middle/src/ty/generic_args.rs @@ -396,14 +396,14 @@ impl<'tcx> GenericArgs<'tcx> { InlineConstArgs { args: self } } - /// Creates an `GenericArgs` that maps each generic parameter to itself. + /// Creates a [`GenericArgs`] that maps each generic parameter to itself. pub fn identity_for_item(tcx: TyCtxt<'tcx>, def_id: impl Into<DefId>) -> GenericArgsRef<'tcx> { Self::for_item(tcx, def_id.into(), |param, _| tcx.mk_param_from_def(param)) } - /// Creates an `GenericArgs` for generic parameter definitions, + /// Creates a [`GenericArgs`] for generic parameter definitions, /// by calling closures to obtain each kind. - /// The closures get to observe the `GenericArgs` as they're + /// The closures get to observe the [`GenericArgs`] as they're /// being built, which can be used to correctly /// replace defaults of generic parameters. pub fn for_item<F>(tcx: TyCtxt<'tcx>, def_id: DefId, mut mk_kind: F) -> GenericArgsRef<'tcx> diff --git a/compiler/rustc_mir_build/src/builder/matches/match_pair.rs b/compiler/rustc_mir_build/src/builder/matches/match_pair.rs index fcfacf1e1f8..9670c1716f5 100644 --- a/compiler/rustc_mir_build/src/builder/matches/match_pair.rs +++ b/compiler/rustc_mir_build/src/builder/matches/match_pair.rs @@ -118,7 +118,7 @@ impl<'tcx> MatchPairTree<'tcx> { let place = place_builder.try_to_place(cx); let mut subpairs = Vec::new(); let test_case = match pattern.kind { - PatKind::Wild | PatKind::Error(_) => None, + PatKind::Missing | PatKind::Wild | PatKind::Error(_) => None, PatKind::Or { ref pats } => Some(TestCase::Or { pats: pats.iter().map(|pat| FlatPat::new(place_builder.clone(), pat, cx)).collect(), diff --git a/compiler/rustc_mir_build/src/builder/matches/mod.rs b/compiler/rustc_mir_build/src/builder/matches/mod.rs index 3acf2a6a2a6..977d4f3e931 100644 --- a/compiler/rustc_mir_build/src/builder/matches/mod.rs +++ b/compiler/rustc_mir_build/src/builder/matches/mod.rs @@ -920,6 +920,7 @@ impl<'a, 'tcx> Builder<'a, 'tcx> { PatKind::Constant { .. } | PatKind::Range { .. } + | PatKind::Missing | PatKind::Wild | PatKind::Never | PatKind::Error(_) => {} diff --git a/compiler/rustc_mir_build/src/check_unsafety.rs b/compiler/rustc_mir_build/src/check_unsafety.rs index b52d0d1e300..b6a856a6eb4 100644 --- a/compiler/rustc_mir_build/src/check_unsafety.rs +++ b/compiler/rustc_mir_build/src/check_unsafety.rs @@ -315,6 +315,7 @@ impl<'a, 'tcx> Visitor<'a, 'tcx> for UnsafetyVisitor<'a, 'tcx> { fn visit_pat(&mut self, pat: &'a Pat<'tcx>) { if self.in_union_destructure { match pat.kind { + PatKind::Missing => unreachable!(), // binding to a variable allows getting stuff out of variable PatKind::Binding { .. } // match is conditional on having this value diff --git a/compiler/rustc_mir_build/src/thir/pattern/mod.rs b/compiler/rustc_mir_build/src/thir/pattern/mod.rs index 7ceeebbb70d..73d60cf4442 100644 --- a/compiler/rustc_mir_build/src/thir/pattern/mod.rs +++ b/compiler/rustc_mir_build/src/thir/pattern/mod.rs @@ -284,6 +284,8 @@ impl<'a, 'tcx> PatCtxt<'a, 'tcx> { let mut span = pat.span; let kind = match pat.kind { + hir::PatKind::Missing => PatKind::Missing, + hir::PatKind::Wild => PatKind::Wild, hir::PatKind::Never => PatKind::Never, diff --git a/compiler/rustc_mir_build/src/thir/print.rs b/compiler/rustc_mir_build/src/thir/print.rs index aa8d2e5897e..37248941e2c 100644 --- a/compiler/rustc_mir_build/src/thir/print.rs +++ b/compiler/rustc_mir_build/src/thir/print.rs @@ -664,6 +664,7 @@ impl<'a, 'tcx> ThirPrinter<'a, 'tcx> { print_indented!(self, "kind: PatKind {", depth_lvl); match pat_kind { + PatKind::Missing => unreachable!(), PatKind::Wild => { print_indented!(self, "Wild", depth_lvl + 1); } diff --git a/compiler/rustc_parse/Cargo.toml b/compiler/rustc_parse/Cargo.toml index c9dcab0c871..6504081f0b9 100644 --- a/compiler/rustc_parse/Cargo.toml +++ b/compiler/rustc_parse/Cargo.toml @@ -6,6 +6,7 @@ edition = "2024" [dependencies] # tidy-alphabetical-start bitflags = "2.4.1" +rustc-literal-escaper = "0.0.2" rustc_ast = { path = "../rustc_ast" } rustc_ast_pretty = { path = "../rustc_ast_pretty" } rustc_data_structures = { path = "../rustc_data_structures" } diff --git a/compiler/rustc_parse/src/lexer/mod.rs b/compiler/rustc_parse/src/lexer/mod.rs index 1d17290e1c7..4935fc03256 100644 --- a/compiler/rustc_parse/src/lexer/mod.rs +++ b/compiler/rustc_parse/src/lexer/mod.rs @@ -6,8 +6,8 @@ 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::unescape::{self, EscapeError, Mode}; use rustc_lexer::{Base, Cursor, DocStyle, LiteralKind, RawStrError}; +use rustc_literal_escaper::{EscapeError, Mode, unescape_mixed, unescape_unicode}; use rustc_session::lint::BuiltinLintDiag; use rustc_session::lint::builtin::{ RUST_2021_PREFIXES_INCOMPATIBLE_SYNTAX, RUST_2024_GUARDED_STRING_INCOMPATIBLE_SYNTAX, @@ -970,9 +970,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> { postfix_len: u32, ) -> (token::LitKind, Symbol) { self.cook_common(kind, mode, start, end, prefix_len, postfix_len, |src, mode, callback| { - unescape::unescape_unicode(src, mode, &mut |span, result| { - callback(span, result.map(drop)) - }) + unescape_unicode(src, mode, &mut |span, result| callback(span, result.map(drop))) }) } @@ -986,9 +984,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> { postfix_len: u32, ) -> (token::LitKind, Symbol) { self.cook_common(kind, mode, start, end, prefix_len, postfix_len, |src, mode, callback| { - unescape::unescape_mixed(src, mode, &mut |span, result| { - callback(span, result.map(drop)) - }) + unescape_mixed(src, mode, &mut |span, result| callback(span, result.map(drop))) }) } } diff --git a/compiler/rustc_parse/src/lexer/unescape_error_reporting.rs b/compiler/rustc_parse/src/lexer/unescape_error_reporting.rs index 2e066f0179c..ec59a1a0131 100644 --- a/compiler/rustc_parse/src/lexer/unescape_error_reporting.rs +++ b/compiler/rustc_parse/src/lexer/unescape_error_reporting.rs @@ -4,7 +4,7 @@ use std::iter::once; use std::ops::Range; use rustc_errors::{Applicability, DiagCtxtHandle, ErrorGuaranteed}; -use rustc_lexer::unescape::{EscapeError, Mode}; +use rustc_literal_escaper::{EscapeError, Mode}; use rustc_span::{BytePos, Span}; use tracing::debug; diff --git a/compiler/rustc_parse/src/parser/expr.rs b/compiler/rustc_parse/src/parser/expr.rs index 841d967d934..9c457f150a3 100644 --- a/compiler/rustc_parse/src/parser/expr.rs +++ b/compiler/rustc_parse/src/parser/expr.rs @@ -21,7 +21,7 @@ use rustc_ast::{ }; use rustc_data_structures::stack::ensure_sufficient_stack; use rustc_errors::{Applicability, Diag, PResult, StashKey, Subdiagnostic}; -use rustc_lexer::unescape::unescape_char; +use rustc_literal_escaper::unescape_char; use rustc_macros::Subdiagnostic; use rustc_session::errors::{ExprParenthesesNeeded, report_lit_error}; use rustc_session::lint::BuiltinLintDiag; diff --git a/compiler/rustc_parse/src/parser/item.rs b/compiler/rustc_parse/src/parser/item.rs index 9405b58ab3b..3647bf2c378 100644 --- a/compiler/rustc_parse/src/parser/item.rs +++ b/compiler/rustc_parse/src/parser/item.rs @@ -2987,9 +2987,7 @@ impl<'a> Parser<'a> { } match ty { Ok(ty) => { - let ident = Ident::new(kw::Empty, this.prev_token.span); - let bm = BindingMode::NONE; - let pat = this.mk_pat_ident(ty.span, bm, ident); + let pat = this.mk_pat(ty.span, PatKind::Missing); (pat, ty) } // If this is a C-variadic argument and we hit an error, return the error. diff --git a/compiler/rustc_parse_format/Cargo.toml b/compiler/rustc_parse_format/Cargo.toml index 289e062fb5e..52f23c00d4b 100644 --- a/compiler/rustc_parse_format/Cargo.toml +++ b/compiler/rustc_parse_format/Cargo.toml @@ -5,6 +5,7 @@ edition = "2024" [dependencies] # tidy-alphabetical-start +rustc-literal-escaper = "0.0.2" rustc_lexer = { path = "../rustc_lexer" } # tidy-alphabetical-end diff --git a/compiler/rustc_parse_format/src/lib.rs b/compiler/rustc_parse_format/src/lib.rs index 97931742985..c59e6cb5c33 100644 --- a/compiler/rustc_parse_format/src/lib.rs +++ b/compiler/rustc_parse_format/src/lib.rs @@ -18,7 +18,7 @@ pub use Alignment::*; pub use Count::*; pub use Position::*; -use rustc_lexer::unescape; +use rustc_literal_escaper::{Mode, unescape_unicode}; // Note: copied from rustc_span /// Range inside of a `Span` used for diagnostics when we only have access to relative positions. @@ -1094,11 +1094,9 @@ fn find_width_map_from_snippet( fn unescape_string(string: &str) -> Option<String> { let mut buf = String::new(); let mut ok = true; - unescape::unescape_unicode(string, unescape::Mode::Str, &mut |_, unescaped_char| { - match unescaped_char { - Ok(c) => buf.push(c), - Err(_) => ok = false, - } + unescape_unicode(string, Mode::Str, &mut |_, unescaped_char| match unescaped_char { + Ok(c) => buf.push(c), + Err(_) => ok = false, }); ok.then_some(buf) diff --git a/compiler/rustc_passes/src/input_stats.rs b/compiler/rustc_passes/src/input_stats.rs index 1278e98afcf..71815448172 100644 --- a/compiler/rustc_passes/src/input_stats.rs +++ b/compiler/rustc_passes/src/input_stats.rs @@ -295,6 +295,7 @@ impl<'v> hir_visit::Visitor<'v> for StatCollector<'v> { record_variants!( (self, p, p.kind, Some(p.hir_id), hir, Pat, PatKind), [ + Missing, Wild, Binding, Struct, @@ -597,6 +598,7 @@ impl<'v> ast_visit::Visitor<'v> for StatCollector<'v> { record_variants!( (self, p, p.kind, None, ast, Pat, PatKind), [ + Missing, Wild, Ident, Struct, diff --git a/compiler/rustc_passes/src/liveness.rs b/compiler/rustc_passes/src/liveness.rs index ed70d9ee91f..06eb76c30c5 100644 --- a/compiler/rustc_passes/src/liveness.rs +++ b/compiler/rustc_passes/src/liveness.rs @@ -96,7 +96,7 @@ use rustc_middle::query::Providers; use rustc_middle::span_bug; use rustc_middle::ty::{self, RootVariableMinCaptureList, Ty, TyCtxt}; use rustc_session::lint; -use rustc_span::{BytePos, Span, Symbol, kw, sym}; +use rustc_span::{BytePos, Span, Symbol, sym}; use tracing::{debug, instrument}; use self::LiveNodeKind::*; @@ -1481,9 +1481,6 @@ impl<'tcx> Liveness<'_, 'tcx> { fn should_warn(&self, var: Variable) -> Option<String> { let name = self.ir.variable_name(var); - if name == kw::Empty { - return None; - } let name = name.as_str(); if name.as_bytes()[0] == b'_' { return None; @@ -1655,7 +1652,7 @@ impl<'tcx> Liveness<'_, 'tcx> { // `&'name Ty` -> `&'name mut Ty` or `&Ty` -> `&mut Ty` Some(mut_ty.ty.span.shrink_to_lo()) }; - let pre = if lt.ident.span.lo() == lt.ident.span.hi() { "" } else { " " }; + let pre = if lt.ident.span.is_empty() { "" } else { " " }; Some(errors::UnusedAssignSuggestion { ty_span, pre, diff --git a/compiler/rustc_pattern_analysis/src/rustc.rs b/compiler/rustc_pattern_analysis/src/rustc.rs index 31c4ee0fa0b..88d45ead295 100644 --- a/compiler/rustc_pattern_analysis/src/rustc.rs +++ b/compiler/rustc_pattern_analysis/src/rustc.rs @@ -460,7 +460,7 @@ impl<'p, 'tcx: 'p> RustcPatCtxt<'p, 'tcx> { PatKind::AscribeUserType { subpattern, .. } | PatKind::ExpandedConstant { subpattern, .. } => return self.lower_pat(subpattern), PatKind::Binding { subpattern: Some(subpat), .. } => return self.lower_pat(subpat), - PatKind::Binding { subpattern: None, .. } | PatKind::Wild => { + PatKind::Missing | PatKind::Binding { subpattern: None, .. } | PatKind::Wild => { ctor = Wildcard; fields = vec![]; arity = 0; diff --git a/compiler/rustc_resolve/src/late.rs b/compiler/rustc_resolve/src/late.rs index 20e19caf909..1389e8c811e 100644 --- a/compiler/rustc_resolve/src/late.rs +++ b/compiler/rustc_resolve/src/late.rs @@ -4009,22 +4009,17 @@ impl<'a, 'ast, 'ra: 'ast, 'tcx> LateResolutionVisitor<'a, 'ast, 'ra, 'tcx> { self.report_error(ident.span, error(ident)); } - // Record as bound if it's valid: - let ident_valid = ident.name != kw::Empty; - if ident_valid { - bindings.last_mut().unwrap().1.insert(ident); - } + // Record as bound. + bindings.last_mut().unwrap().1.insert(ident); if already_bound_or { // `Variant1(a) | Variant2(a)`, ok // Reuse definition from the first `a`. self.innermost_rib_bindings(ValueNS)[&ident] } else { + // A completely fresh binding is added to the set. let res = Res::Local(pat_id); - if ident_valid { - // A completely fresh binding add to the set if it's valid. - self.innermost_rib_bindings(ValueNS).insert(ident, res); - } + self.innermost_rib_bindings(ValueNS).insert(ident, res); res } } diff --git a/compiler/rustc_type_ir/Cargo.toml b/compiler/rustc_type_ir/Cargo.toml index 4adf7157926..83d3d78298e 100644 --- a/compiler/rustc_type_ir/Cargo.toml +++ b/compiler/rustc_type_ir/Cargo.toml @@ -7,6 +7,7 @@ edition = "2024" # tidy-alphabetical-start bitflags = "2.4.1" derive-where = "1.2.7" +ena = "0.14.3" indexmap = "2.0.0" rustc-hash = "1.1.0" rustc_ast_ir = { path = "../rustc_ast_ir", default-features = false } diff --git a/compiler/rustc_type_ir/src/data_structures/mod.rs b/compiler/rustc_type_ir/src/data_structures/mod.rs index 30c67d10d0e..a72669cbd18 100644 --- a/compiler/rustc_type_ir/src/data_structures/mod.rs +++ b/compiler/rustc_type_ir/src/data_structures/mod.rs @@ -1,5 +1,6 @@ use std::hash::BuildHasherDefault; +pub use ena::unify::{NoError, UnifyKey, UnifyValue}; use rustc_hash::FxHasher; pub use rustc_hash::{FxHashMap as HashMap, FxHashSet as HashSet}; diff --git a/compiler/rustc_type_ir/src/ty_kind.rs b/compiler/rustc_type_ir/src/ty_kind.rs index d35b22d517c..753a72a051a 100644 --- a/compiler/rustc_type_ir/src/ty_kind.rs +++ b/compiler/rustc_type_ir/src/ty_kind.rs @@ -6,9 +6,8 @@ use rustc_ast_ir::Mutability; #[cfg(feature = "nightly")] use rustc_data_structures::stable_hasher::{HashStable, StableHasher}; #[cfg(feature = "nightly")] -use rustc_data_structures::unify::{NoError, UnifyKey, UnifyValue}; -#[cfg(feature = "nightly")] use rustc_macros::{Decodable_NoContext, Encodable_NoContext, HashStable_NoContext}; +use rustc_type_ir::data_structures::{NoError, UnifyKey, UnifyValue}; use rustc_type_ir_macros::{Lift_Generic, TypeFoldable_Generic, TypeVisitable_Generic}; use self::TyKind::*; @@ -796,7 +795,6 @@ pub enum InferTy { /// Raw `TyVid` are used as the unification key for `sub_relations`; /// they carry no values. -#[cfg(feature = "nightly")] impl UnifyKey for TyVid { type Value = (); #[inline] @@ -812,7 +810,6 @@ impl UnifyKey for TyVid { } } -#[cfg(feature = "nightly")] impl UnifyValue for IntVarValue { type Error = NoError; @@ -832,7 +829,6 @@ impl UnifyValue for IntVarValue { } } -#[cfg(feature = "nightly")] impl UnifyKey for IntVid { type Value = IntVarValue; #[inline] // make this function eligible for inlining - it is quite hot. @@ -848,7 +844,6 @@ impl UnifyKey for IntVid { } } -#[cfg(feature = "nightly")] impl UnifyValue for FloatVarValue { type Error = NoError; @@ -866,7 +861,6 @@ impl UnifyValue for FloatVarValue { } } -#[cfg(feature = "nightly")] impl UnifyKey for FloatVid { type Value = FloatVarValue; #[inline] diff --git a/library/Cargo.lock b/library/Cargo.lock index 23d9d926eba..ad634e9f794 100644 --- a/library/Cargo.lock +++ b/library/Cargo.lock @@ -176,9 +176,9 @@ dependencies = [ [[package]] name = "miniz_oxide" -version = "0.8.5" +version = "0.8.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8e3e04debbb59698c15bacbb6d93584a8c0ca9cc3213cb423d31f760d8843ce5" +checksum = "ff70ce3e48ae43fa075863cef62e8b43b71a4f2382229920e0df362592919430" dependencies = [ "adler2", "compiler_builtins", @@ -235,6 +235,7 @@ name = "proc_macro" version = "0.0.0" dependencies = [ "core", + "rustc-literal-escaper", "std", ] @@ -311,6 +312,15 @@ dependencies = [ ] [[package]] +name = "rustc-literal-escaper" +version = "0.0.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0041b6238913c41fe704213a4a9329e2f685a156d1781998128b4149c230ad04" +dependencies = [ + "rustc-std-workspace-std", +] + +[[package]] name = "rustc-std-workspace-alloc" version = "1.99.0" dependencies = [ diff --git a/library/alloc/src/lib.rs b/library/alloc/src/lib.rs index f0cdb1e4e0f..04858667230 100644 --- a/library/alloc/src/lib.rs +++ b/library/alloc/src/lib.rs @@ -135,6 +135,7 @@ #![feature(pattern)] #![feature(pin_coerce_unsized_trait)] #![feature(pointer_like_trait)] +#![feature(ptr_alignment_type)] #![feature(ptr_internals)] #![feature(ptr_metadata)] #![feature(set_ptr_value)] diff --git a/library/alloc/src/raw_vec/mod.rs b/library/alloc/src/raw_vec/mod.rs index 99ebc5c4bfc..a989e5b55b3 100644 --- a/library/alloc/src/raw_vec/mod.rs +++ b/library/alloc/src/raw_vec/mod.rs @@ -6,7 +6,7 @@ use core::marker::PhantomData; use core::mem::{ManuallyDrop, MaybeUninit, SizedTypeProperties}; -use core::ptr::{self, NonNull, Unique}; +use core::ptr::{self, Alignment, NonNull, Unique}; use core::{cmp, hint}; #[cfg(not(no_global_oom_handling))] @@ -177,7 +177,7 @@ impl<T, A: Allocator> RawVec<T, A> { /// the returned `RawVec`. #[inline] pub(crate) const fn new_in(alloc: A) -> Self { - Self { inner: RawVecInner::new_in(alloc, align_of::<T>()), _marker: PhantomData } + Self { inner: RawVecInner::new_in(alloc, Alignment::of::<T>()), _marker: PhantomData } } /// Like `with_capacity`, but parameterized over the choice of @@ -409,8 +409,8 @@ unsafe impl<#[may_dangle] T, A: Allocator> Drop for RawVec<T, A> { impl<A: Allocator> RawVecInner<A> { #[inline] - const fn new_in(alloc: A, align: usize) -> Self { - let ptr = unsafe { core::mem::transmute(align) }; + const fn new_in(alloc: A, align: Alignment) -> Self { + let ptr = Unique::from_non_null(NonNull::without_provenance(align.as_nonzero())); // `cap: 0` means "unallocated". zero-sized types are ignored. Self { ptr, cap: ZERO_CAP, alloc } } @@ -465,7 +465,7 @@ impl<A: Allocator> RawVecInner<A> { // Don't allocate here because `Drop` will not deallocate when `capacity` is 0. if layout.size() == 0 { - return Ok(Self::new_in(alloc, elem_layout.align())); + return Ok(Self::new_in(alloc, elem_layout.alignment())); } if let Err(err) = alloc_guard(layout.size()) { diff --git a/library/alloctests/lib.rs b/library/alloctests/lib.rs index 6ce8a6d9ca1..56e60ed4c84 100644 --- a/library/alloctests/lib.rs +++ b/library/alloctests/lib.rs @@ -28,6 +28,8 @@ #![feature(iter_next_chunk)] #![feature(maybe_uninit_slice)] #![feature(maybe_uninit_uninit_array_transpose)] +#![feature(nonnull_provenance)] +#![feature(ptr_alignment_type)] #![feature(ptr_internals)] #![feature(sized_type_properties)] #![feature(slice_iter_mut_as_mut_slice)] diff --git a/library/core/src/alloc/layout.rs b/library/core/src/alloc/layout.rs index 1595a3af883..e8a03aadc33 100644 --- a/library/core/src/alloc/layout.rs +++ b/library/core/src/alloc/layout.rs @@ -520,6 +520,14 @@ impl Layout { unsafe { Ok(Layout::from_size_align_unchecked(array_size, align.as_usize())) } } } + + /// Perma-unstable access to `align` as `Alignment` type. + #[unstable(issue = "none", feature = "std_internals")] + #[doc(hidden)] + #[inline] + pub const fn alignment(&self) -> Alignment { + self.align + } } #[stable(feature = "alloc_layout", since = "1.28.0")] diff --git a/library/core/src/ptr/unique.rs b/library/core/src/ptr/unique.rs index 4810ebe01f9..d688ce2a07a 100644 --- a/library/core/src/ptr/unique.rs +++ b/library/core/src/ptr/unique.rs @@ -100,6 +100,12 @@ impl<T: ?Sized> Unique<T> { } } + /// Create a new `Unique` from a `NonNull` in const context. + #[inline] + pub const fn from_non_null(pointer: NonNull<T>) -> Self { + Unique { pointer, _marker: PhantomData } + } + /// Acquires the underlying `*mut` pointer. #[must_use = "`self` will be dropped if the result is not used"] #[inline] @@ -202,6 +208,6 @@ impl<T: ?Sized> From<NonNull<T>> for Unique<T> { /// This conversion is infallible since `NonNull` cannot be null. #[inline] fn from(pointer: NonNull<T>) -> Self { - Unique { pointer, _marker: PhantomData } + Unique::from_non_null(pointer) } } diff --git a/library/core/src/slice/iter.rs b/library/core/src/slice/iter.rs index d48248749c2..f507ee563ac 100644 --- a/library/core/src/slice/iter.rs +++ b/library/core/src/slice/iter.rs @@ -1380,14 +1380,16 @@ impl<'a, T> Iterator for Windows<'a, T> { #[inline] fn nth(&mut self, n: usize) -> Option<Self::Item> { - let (end, overflow) = self.size.get().overflowing_add(n); - if end > self.v.len() || overflow { - self.v = &[]; - None - } else { - let nth = &self.v[n..end]; - self.v = &self.v[n + 1..]; + let size = self.size.get(); + if let Some(rest) = self.v.get(n..) + && let Some(nth) = rest.get(..size) + { + self.v = &rest[1..]; Some(nth) + } else { + // setting length to 0 is cheaper than overwriting the pointer when assigning &[] + self.v = &self.v[..0]; // cheaper than &[] + None } } @@ -1427,7 +1429,7 @@ impl<'a, T> DoubleEndedIterator for Windows<'a, T> { fn nth_back(&mut self, n: usize) -> Option<Self::Item> { let (end, overflow) = self.v.len().overflowing_sub(n); if end < self.size.get() || overflow { - self.v = &[]; + self.v = &self.v[..0]; // cheaper than &[] None } else { let ret = &self.v[end - self.size.get()..end]; @@ -1536,17 +1538,15 @@ impl<'a, T> Iterator for Chunks<'a, T> { #[inline] fn nth(&mut self, n: usize) -> Option<Self::Item> { let (start, overflow) = n.overflowing_mul(self.chunk_size); - if start >= self.v.len() || overflow { - self.v = &[]; - None - } else { - let end = match start.checked_add(self.chunk_size) { - Some(sum) => cmp::min(self.v.len(), sum), - None => self.v.len(), - }; - let nth = &self.v[start..end]; - self.v = &self.v[end..]; + // min(len) makes a wrong start harmless, but enables optimizing this to brachless code + let chunk_start = &self.v[start.min(self.v.len())..]; + let (nth, remainder) = chunk_start.split_at(self.chunk_size.min(chunk_start.len())); + if !overflow && start < self.v.len() { + self.v = remainder; Some(nth) + } else { + self.v = &self.v[..0]; // cheaper than &[] + None } } @@ -1609,7 +1609,7 @@ impl<'a, T> DoubleEndedIterator for Chunks<'a, T> { fn nth_back(&mut self, n: usize) -> Option<Self::Item> { let len = self.len(); if n >= len { - self.v = &[]; + self.v = &self.v[..0]; // cheaper than &[] None } else { let start = (len - 1 - n) * self.chunk_size; @@ -1933,7 +1933,7 @@ impl<'a, T> Iterator for ChunksExact<'a, T> { fn nth(&mut self, n: usize) -> Option<Self::Item> { let (start, overflow) = n.overflowing_mul(self.chunk_size); if start >= self.v.len() || overflow { - self.v = &[]; + self.v = &self.v[..0]; // cheaper than &[] None } else { let (_, snd) = self.v.split_at(start); @@ -1971,7 +1971,7 @@ impl<'a, T> DoubleEndedIterator for ChunksExact<'a, T> { fn nth_back(&mut self, n: usize) -> Option<Self::Item> { let len = self.len(); if n >= len { - self.v = &[]; + self.v = &self.v[..0]; // cheaper than &[] None } else { let start = (len - 1 - n) * self.chunk_size; @@ -2638,7 +2638,7 @@ impl<'a, T> Iterator for RChunks<'a, T> { fn nth(&mut self, n: usize) -> Option<Self::Item> { let (end, overflow) = n.overflowing_mul(self.chunk_size); if end >= self.v.len() || overflow { - self.v = &[]; + self.v = &self.v[..0]; // cheaper than &[] None } else { // Can't underflow because of the check above @@ -2695,7 +2695,7 @@ impl<'a, T> DoubleEndedIterator for RChunks<'a, T> { fn nth_back(&mut self, n: usize) -> Option<Self::Item> { let len = self.len(); if n >= len { - self.v = &[]; + self.v = &self.v[..0]; // cheaper than &[] None } else { // can't underflow because `n < len` @@ -3023,7 +3023,7 @@ impl<'a, T> Iterator for RChunksExact<'a, T> { fn nth(&mut self, n: usize) -> Option<Self::Item> { let (end, overflow) = n.overflowing_mul(self.chunk_size); if end >= self.v.len() || overflow { - self.v = &[]; + self.v = &self.v[..0]; // cheaper than &[] None } else { let (fst, _) = self.v.split_at(self.v.len() - end); @@ -3062,7 +3062,7 @@ impl<'a, T> DoubleEndedIterator for RChunksExact<'a, T> { fn nth_back(&mut self, n: usize) -> Option<Self::Item> { let len = self.len(); if n >= len { - self.v = &[]; + self.v = &self.v[..0]; // cheaper than &[] None } else { // now that we know that `n` corresponds to a chunk, diff --git a/library/core/src/slice/mod.rs b/library/core/src/slice/mod.rs index e54840c8fcc..c7e8d5f989d 100644 --- a/library/core/src/slice/mod.rs +++ b/library/core/src/slice/mod.rs @@ -8,7 +8,7 @@ use crate::cmp::Ordering::{self, Equal, Greater, Less}; use crate::intrinsics::{exact_div, unchecked_sub}; -use crate::mem::{self, SizedTypeProperties}; +use crate::mem::{self, MaybeUninit, SizedTypeProperties}; use crate::num::NonZero; use crate::ops::{OneSidedRange, OneSidedRangeBound, Range, RangeBounds, RangeInclusive}; use crate::panic::const_panic; @@ -4579,7 +4579,7 @@ impl<T> [T] { // or generate worse code otherwise. This is also why we need to go // through a raw pointer here. let slice: *mut [T] = self; - let mut arr: mem::MaybeUninit<[&mut I::Output; N]> = mem::MaybeUninit::uninit(); + let mut arr: MaybeUninit<[&mut I::Output; N]> = MaybeUninit::uninit(); let arr_ptr = arr.as_mut_ptr(); // SAFETY: We expect `indices` to contain disjunct values that are @@ -4764,6 +4764,55 @@ impl<T> [T] { } } +impl<T> [MaybeUninit<T>] { + /// Transmutes the mutable uninitialized slice to a mutable uninitialized slice of + /// another type, ensuring alignment of the types is maintained. + /// + /// This is a safe wrapper around [`slice::align_to_mut`], so inherits the same + /// guarantees as that method. + /// + /// # Examples + /// + /// ``` + /// #![feature(align_to_uninit_mut)] + /// use std::mem::MaybeUninit; + /// + /// pub struct BumpAllocator<'scope> { + /// memory: &'scope mut [MaybeUninit<u8>], + /// } + /// + /// impl<'scope> BumpAllocator<'scope> { + /// pub fn new(memory: &'scope mut [MaybeUninit<u8>]) -> Self { + /// Self { memory } + /// } + /// pub fn try_alloc_uninit<T>(&mut self) -> Option<&'scope mut MaybeUninit<T>> { + /// let first_end = self.memory.as_ptr().align_offset(align_of::<T>()) + size_of::<T>(); + /// let prefix = self.memory.split_off_mut(..first_end)?; + /// Some(&mut prefix.align_to_uninit_mut::<T>().1[0]) + /// } + /// pub fn try_alloc_u32(&mut self, value: u32) -> Option<&'scope mut u32> { + /// let uninit = self.try_alloc_uninit()?; + /// Some(uninit.write(value)) + /// } + /// } + /// + /// let mut memory = [MaybeUninit::<u8>::uninit(); 10]; + /// let mut allocator = BumpAllocator::new(&mut memory); + /// let v = allocator.try_alloc_u32(42); + /// assert_eq!(v, Some(&mut 42)); + /// ``` + #[unstable(feature = "align_to_uninit_mut", issue = "139062")] + #[inline] + #[must_use] + pub fn align_to_uninit_mut<U>(&mut self) -> (&mut Self, &mut [MaybeUninit<U>], &mut Self) { + // SAFETY: `MaybeUninit` is transparent. Correct size and alignment are guaranteed by + // `align_to_mut` itself. Therefore the only thing that we have to ensure for a safe + // `transmute` is that the values are valid for the types involved. But for `MaybeUninit` + // any values are valid, so this operation is safe. + unsafe { self.align_to_mut() } + } +} + impl<T, const N: usize> [[T; N]] { /// Takes a `&[[T; N]]`, and flattens it to a `&[T]`. /// diff --git a/library/proc_macro/Cargo.toml b/library/proc_macro/Cargo.toml index 72cb4e4166f..b8bc2a3af4c 100644 --- a/library/proc_macro/Cargo.toml +++ b/library/proc_macro/Cargo.toml @@ -9,3 +9,4 @@ std = { path = "../std" } # `core` when resolving doc links. Without this line a different `core` will be # loaded from sysroot causing duplicate lang items and other similar errors. core = { path = "../core" } +rustc-literal-escaper = { version = "0.0.2", features = ["rustc-dep-of-std"] } diff --git a/library/proc_macro/src/lib.rs b/library/proc_macro/src/lib.rs index d9141eab591..f1cf0c5a2db 100644 --- a/library/proc_macro/src/lib.rs +++ b/library/proc_macro/src/lib.rs @@ -27,6 +27,7 @@ #![feature(panic_can_unwind)] #![feature(restricted_std)] #![feature(rustc_attrs)] +#![feature(stmt_expr_attributes)] #![feature(extend_one)] #![recursion_limit = "256"] #![allow(internal_features)] @@ -51,11 +52,24 @@ use std::{error, fmt}; #[unstable(feature = "proc_macro_diagnostic", issue = "54140")] pub use diagnostic::{Diagnostic, Level, MultiSpan}; +#[unstable(feature = "proc_macro_value", issue = "136652")] +pub use rustc_literal_escaper::EscapeError; +use rustc_literal_escaper::{MixedUnit, Mode, byte_from_char, unescape_mixed, unescape_unicode}; #[unstable(feature = "proc_macro_totokens", issue = "130977")] pub use to_tokens::ToTokens; use crate::escape::{EscapeOptions, escape_bytes}; +/// Errors returned when trying to retrieve a literal unescaped value. +#[unstable(feature = "proc_macro_value", issue = "136652")] +#[derive(Debug, PartialEq, Eq)] +pub enum ConversionErrorKind { + /// The literal failed to be escaped, take a look at [`EscapeError`] for more information. + FailedToUnescape(EscapeError), + /// Trying to convert a literal with the wrong type. + InvalidLiteralKind, +} + /// Determines whether proc_macro has been made accessible to the currently /// running program. /// @@ -1451,6 +1465,107 @@ impl Literal { } }) } + + /// Returns the unescaped string value if the current literal is a string or a string literal. + #[unstable(feature = "proc_macro_value", issue = "136652")] + pub fn str_value(&self) -> Result<String, ConversionErrorKind> { + self.0.symbol.with(|symbol| match self.0.kind { + bridge::LitKind::Str => { + if symbol.contains('\\') { + let mut buf = String::with_capacity(symbol.len()); + let mut error = None; + // Force-inlining here is aggressive but the closure is + // called on every char in the string, so it can be hot in + // programs with many long strings containing escapes. + unescape_unicode( + symbol, + Mode::Str, + &mut #[inline(always)] + |_, c| match c { + Ok(c) => buf.push(c), + Err(err) => { + if err.is_fatal() { + error = Some(ConversionErrorKind::FailedToUnescape(err)); + } + } + }, + ); + if let Some(error) = error { Err(error) } else { Ok(buf) } + } else { + Ok(symbol.to_string()) + } + } + bridge::LitKind::StrRaw(_) => Ok(symbol.to_string()), + _ => Err(ConversionErrorKind::InvalidLiteralKind), + }) + } + + /// Returns the unescaped string value if the current literal is a c-string or a c-string + /// literal. + #[unstable(feature = "proc_macro_value", issue = "136652")] + pub fn cstr_value(&self) -> Result<Vec<u8>, ConversionErrorKind> { + self.0.symbol.with(|symbol| match self.0.kind { + bridge::LitKind::CStr => { + let mut error = None; + let mut buf = Vec::with_capacity(symbol.len()); + + unescape_mixed(symbol, Mode::CStr, &mut |_span, c| match c { + Ok(MixedUnit::Char(c)) => { + buf.extend_from_slice(c.encode_utf8(&mut [0; 4]).as_bytes()) + } + Ok(MixedUnit::HighByte(b)) => buf.push(b), + Err(err) => { + if err.is_fatal() { + error = Some(ConversionErrorKind::FailedToUnescape(err)); + } + } + }); + if let Some(error) = error { + Err(error) + } else { + buf.push(0); + Ok(buf) + } + } + bridge::LitKind::CStrRaw(_) => { + // Raw strings have no escapes so we can convert the symbol + // directly to a `Lrc<u8>` after appending the terminating NUL + // char. + let mut buf = symbol.to_owned().into_bytes(); + buf.push(0); + Ok(buf) + } + _ => Err(ConversionErrorKind::InvalidLiteralKind), + }) + } + + /// Returns the unescaped string value if the current literal is a byte string or a byte string + /// literal. + #[unstable(feature = "proc_macro_value", issue = "136652")] + pub fn byte_str_value(&self) -> Result<Vec<u8>, ConversionErrorKind> { + self.0.symbol.with(|symbol| match self.0.kind { + bridge::LitKind::ByteStr => { + let mut buf = Vec::with_capacity(symbol.len()); + let mut error = None; + + unescape_unicode(symbol, Mode::ByteStr, &mut |_, c| match c { + Ok(c) => buf.push(byte_from_char(c)), + Err(err) => { + if err.is_fatal() { + error = Some(ConversionErrorKind::FailedToUnescape(err)); + } + } + }); + if let Some(error) = error { Err(error) } else { Ok(buf) } + } + bridge::LitKind::ByteStrRaw(_) => { + // Raw strings have no escapes so we can convert the symbol + // directly to a `Lrc<u8>`. + Ok(symbol.to_owned().into_bytes()) + } + _ => Err(ConversionErrorKind::InvalidLiteralKind), + }) + } } /// Parse a single literal from its stringified representation. diff --git a/library/std/src/sys/io/io_slice/iovec.rs b/library/std/src/sys/io/io_slice/iovec.rs index 072191315f7..df56358969a 100644 --- a/library/std/src/sys/io/io_slice/iovec.rs +++ b/library/std/src/sys/io/io_slice/iovec.rs @@ -1,6 +1,6 @@ #[cfg(target_os = "hermit")] use hermit_abi::iovec; -#[cfg(target_family = "unix")] +#[cfg(any(target_family = "unix", target_os = "trusty"))] use libc::iovec; use crate::ffi::c_void; diff --git a/library/std/src/sys/io/mod.rs b/library/std/src/sys/io/mod.rs index e00b479109f..4d0365d42fd 100644 --- a/library/std/src/sys/io/mod.rs +++ b/library/std/src/sys/io/mod.rs @@ -2,7 +2,7 @@ mod io_slice { cfg_if::cfg_if! { - if #[cfg(any(target_family = "unix", target_os = "hermit", target_os = "solid_asp3"))] { + if #[cfg(any(target_family = "unix", target_os = "hermit", target_os = "solid_asp3", target_os = "trusty"))] { mod iovec; pub use iovec::*; } else if #[cfg(target_os = "windows")] { diff --git a/library/std/src/sys/stdio/trusty.rs b/library/std/src/sys/stdio/trusty.rs index d393e95394d..e05461aa44a 100644 --- a/library/std/src/sys/stdio/trusty.rs +++ b/library/std/src/sys/stdio/trusty.rs @@ -1,21 +1,14 @@ -use crate::io; +#[expect(dead_code)] +#[path = "unsupported.rs"] +mod unsupported_stdio; -pub struct Stdin; +use crate::cmp; +use crate::io::{self, IoSlice}; + +pub type Stdin = unsupported_stdio::Stdin; pub struct Stdout; pub struct Stderr; -impl Stdin { - pub const fn new() -> Stdin { - Stdin - } -} - -impl io::Read for Stdin { - fn read(&mut self, _buf: &mut [u8]) -> io::Result<usize> { - Ok(0) - } -} - impl Stdout { pub const fn new() -> Stdout { Stdout @@ -24,7 +17,16 @@ impl Stdout { impl io::Write for Stdout { fn write(&mut self, buf: &[u8]) -> io::Result<usize> { - _write(libc::STDOUT_FILENO, buf) + write(libc::STDOUT_FILENO, buf) + } + + fn write_vectored(&mut self, bufs: &[IoSlice<'_>]) -> io::Result<usize> { + write_vectored(libc::STDOUT_FILENO, bufs) + } + + #[inline] + fn is_write_vectored(&self) -> bool { + true } fn flush(&mut self) -> io::Result<()> { @@ -40,7 +42,16 @@ impl Stderr { impl io::Write for Stderr { fn write(&mut self, buf: &[u8]) -> io::Result<usize> { - _write(libc::STDERR_FILENO, buf) + write(libc::STDERR_FILENO, buf) + } + + fn write_vectored(&mut self, bufs: &[IoSlice<'_>]) -> io::Result<usize> { + write_vectored(libc::STDERR_FILENO, bufs) + } + + #[inline] + fn is_write_vectored(&self) -> bool { + true } fn flush(&mut self) -> io::Result<()> { @@ -48,7 +59,7 @@ impl io::Write for Stderr { } } -pub const STDIN_BUF_SIZE: usize = 0; +pub const STDIN_BUF_SIZE: usize = unsupported_stdio::STDIN_BUF_SIZE; pub fn is_ebadf(_err: &io::Error) -> bool { true @@ -58,24 +69,24 @@ pub fn panic_output() -> Option<impl io::Write> { Some(Stderr) } -fn _write(fd: i32, message: &[u8]) -> io::Result<usize> { - let mut iov = libc::iovec { iov_base: message.as_ptr() as *mut _, iov_len: message.len() }; - loop { - // SAFETY: syscall, safe arguments. - let ret = unsafe { libc::writev(fd, &iov, 1) }; - if ret < 0 { - return Err(io::Error::last_os_error()); - } - let ret = ret as usize; - if ret > iov.iov_len { - return Err(io::Error::last_os_error()); - } - if ret == iov.iov_len { - return Ok(message.len()); - } - // SAFETY: ret has been checked to be less than the length of - // the buffer - iov.iov_base = unsafe { iov.iov_base.add(ret) }; - iov.iov_len -= ret; +fn write(fd: i32, buf: &[u8]) -> io::Result<usize> { + let iov = libc::iovec { iov_base: buf.as_ptr() as *mut _, iov_len: buf.len() }; + // SAFETY: syscall, safe arguments. + let ret = unsafe { libc::writev(fd, &iov, 1) }; + // This check includes ret < 0, since the length is at most isize::MAX. + if ret as usize > iov.iov_len { + return Err(io::Error::last_os_error()); + } + Ok(ret as usize) +} + +fn write_vectored(fd: i32, bufs: &[IoSlice<'_>]) -> io::Result<usize> { + let iov = bufs.as_ptr() as *const libc::iovec; + let len = cmp::min(bufs.len(), libc::c_int::MAX as usize) as libc::c_int; + // SAFETY: syscall, safe arguments. + let ret = unsafe { libc::writev(fd, iov, len) }; + if ret < 0 { + return Err(io::Error::last_os_error()); } + Ok(ret as usize) } diff --git a/src/bootstrap/src/lib.rs b/src/bootstrap/src/lib.rs index 843d474f92d..7c60e042142 100644 --- a/src/bootstrap/src/lib.rs +++ b/src/bootstrap/src/lib.rs @@ -748,7 +748,7 @@ impl Build { features.push("llvm"); } // keep in sync with `bootstrap/compile.rs:rustc_cargo_env` - if self.config.rust_randomize_layout { + if self.config.rust_randomize_layout && check("rustc_randomized_layouts") { features.push("rustc_randomized_layouts"); } diff --git a/src/doc/style-guide/src/expressions.md b/src/doc/style-guide/src/expressions.md index 12037b5992e..031e59d86e1 100644 --- a/src/doc/style-guide/src/expressions.md +++ b/src/doc/style-guide/src/expressions.md @@ -521,8 +521,11 @@ self.pre_comment.as_ref().map_or( ## Control flow expressions -This section covers `if`, `if let`, `loop`, `while`, `while let`, and `for` -expressions. +This section covers `for` and `loop` expressions, as well as `if` and `while` +expressions with their sub-expression variants. This includes those with a +single `let` sub-expression (i.e. `if let` and `while let`) +as well as "let-chains": those with one or more `let` sub-expressions and +one or more bool-type conditions (i.e. `if a && let Some(b) = c`). Put the keyword, any initial clauses, and the opening brace of the block all on a single line, if they fit. Apply the usual rules for [block @@ -548,10 +551,11 @@ if let ... { } ``` -If the control line needs to be broken, prefer to break before the `=` in `* -let` expressions and before `in` in a `for` expression; block-indent the -following line. If the control line is broken for any reason, put the opening -brace on its own line, not indented. Examples: +If the control line needs to be broken, then prefer breaking after the `=` for any +`let` sub-expression in an `if` or `while` expression that does not fit, +and before `in` in a `for` expression; the following line should be block indented. +If the control line is broken for any reason, then the opening brace should be on its +own line and not indented. Examples: ```rust while let Some(foo) @@ -572,6 +576,68 @@ if a_long_expression { ... } + +if let Some(a) = b + && another_long_expression + && a_third_long_expression +{ + // ... +} + +if let Some(relatively_long_thing) + = a_long_expression + && another_long_expression + && a_third_long_expression +{ + // ... +} + +if some_expr + && another_long_expression + && let Some(relatively_long_thing) = + a_long_long_long_long_long_long_really_reallllllllllyyyyyyy_long_expression + && a_third_long_expression +{ + // ... +} +``` + +A let-chain control line is allowed to be formatted on a single line provided +it only consists of two clauses, with the first, left-hand side operand being a literal or an +`ident` (which can optionally be preceded by any number of unary prefix operators), +and the second, right-hand side operand being a single-line `let` clause. Otherwise, +the control line must be broken and formatted according to the above rules. For example: + +```rust +if a && let Some(b) = foo() { + // ... +} + +if true && let Some(b) = foo() { + // ... +} + +let operator = if !from_hir_call && let Some(p) = parent { + // ... +}; + +if let Some(b) = foo() + && a +{ + // .. +} + +if foo() + && let Some(b) = bar +{ + // ... +} + +if gen_pos != GenericArgPosition::Type + && let Some(b) = gen_args.bindings.first() +{ + // .. +} ``` Where the initial clause spans multiple lines and ends with one or more closing diff --git a/src/librustdoc/clean/utils.rs b/src/librustdoc/clean/utils.rs index f81db58950c..afcca81a485 100644 --- a/src/librustdoc/clean/utils.rs +++ b/src/librustdoc/clean/utils.rs @@ -304,6 +304,7 @@ pub(crate) fn name_from_pat(p: &hir::Pat<'_>) -> Symbol { Symbol::intern(&match &p.kind { // FIXME(never_patterns): does this make sense? + PatKind::Missing => unreachable!(), PatKind::Wild | PatKind::Err(_) | PatKind::Never diff --git a/src/tools/cargo b/src/tools/cargo -Subproject a6c604d1b8a2f2a8ff1f3ba6092f9fda42f4b7e +Subproject 0e93c5bf6a1d5ee7bc2af63d1afb16cd2879360 diff --git a/src/tools/clippy/clippy_lints/src/equatable_if_let.rs b/src/tools/clippy/clippy_lints/src/equatable_if_let.rs index cd9ab2764ac..3afb687040f 100644 --- a/src/tools/clippy/clippy_lints/src/equatable_if_let.rs +++ b/src/tools/clippy/clippy_lints/src/equatable_if_let.rs @@ -45,6 +45,7 @@ fn unary_pattern(pat: &Pat<'_>) -> bool { pats.iter().all(unary_pattern) } match &pat.kind { + PatKind::Missing => unreachable!(), PatKind::Slice(_, _, _) | PatKind::Range(_, _, _) | PatKind::Binding(..) diff --git a/src/tools/clippy/clippy_lints/src/matches/match_same_arms.rs b/src/tools/clippy/clippy_lints/src/matches/match_same_arms.rs index 250f17fa902..a21597ffb93 100644 --- a/src/tools/clippy/clippy_lints/src/matches/match_same_arms.rs +++ b/src/tools/clippy/clippy_lints/src/matches/match_same_arms.rs @@ -253,6 +253,7 @@ fn iter_matching_struct_fields<'a>( impl<'a> NormalizedPat<'a> { fn from_pat(cx: &LateContext<'_>, arena: &'a DroplessArena, pat: &'a Pat<'_>) -> Self { match pat.kind { + PatKind::Missing => unreachable!(), PatKind::Wild | PatKind::Binding(.., None) => Self::Wild, PatKind::Binding(.., Some(pat)) | PatKind::Box(pat) diff --git a/src/tools/clippy/clippy_lints/src/matches/single_match.rs b/src/tools/clippy/clippy_lints/src/matches/single_match.rs index 56fbd626eef..836c46240ce 100644 --- a/src/tools/clippy/clippy_lints/src/matches/single_match.rs +++ b/src/tools/clippy/clippy_lints/src/matches/single_match.rs @@ -406,6 +406,7 @@ impl<'a> PatState<'a> { pats.iter().map(|p| p.pat), ), + PatKind::Missing => unreachable!(), PatKind::Wild | PatKind::Binding(_, _, _, None) | PatKind::Expr(_) diff --git a/src/tools/clippy/clippy_lints/src/unnested_or_patterns.rs b/src/tools/clippy/clippy_lints/src/unnested_or_patterns.rs index f43715d6752..8966e6851ac 100644 --- a/src/tools/clippy/clippy_lints/src/unnested_or_patterns.rs +++ b/src/tools/clippy/clippy_lints/src/unnested_or_patterns.rs @@ -224,6 +224,7 @@ fn transform_with_focus_on_idx(alternatives: &mut ThinVec<P<Pat>>, focus_idx: us // We're trying to find whatever kind (~"constructor") we found in `alternatives[start..]`. let changed = match &mut focus_kind { + Missing => unreachable!(), // These pattern forms are "leafs" and do not have sub-patterns. // Therefore they are not some form of constructor `C`, // with which a pattern `C(p_0)` may be formed, diff --git a/src/tools/clippy/clippy_lints/src/utils/author.rs b/src/tools/clippy/clippy_lints/src/utils/author.rs index 4309cd2c9ab..b7dcd2ffb0e 100644 --- a/src/tools/clippy/clippy_lints/src/utils/author.rs +++ b/src/tools/clippy/clippy_lints/src/utils/author.rs @@ -676,6 +676,7 @@ impl<'a, 'tcx> PrintVisitor<'a, 'tcx> { } match pat.value.kind { + PatKind::Missing => unreachable!(), PatKind::Wild => kind!("Wild"), PatKind::Never => kind!("Never"), PatKind::Binding(ann, _, name, sub) => { diff --git a/src/tools/clippy/clippy_utils/src/ast_utils/mod.rs b/src/tools/clippy/clippy_utils/src/ast_utils/mod.rs index eba576392eb..c5dce26143b 100644 --- a/src/tools/clippy/clippy_utils/src/ast_utils/mod.rs +++ b/src/tools/clippy/clippy_utils/src/ast_utils/mod.rs @@ -33,6 +33,7 @@ pub fn eq_id(l: Ident, r: Ident) -> bool { pub fn eq_pat(l: &Pat, r: &Pat) -> bool { use PatKind::*; match (&l.kind, &r.kind) { + (Missing, _) | (_, Missing) => unreachable!(), (Paren(l), _) => eq_pat(l, r), (_, Paren(r)) => eq_pat(l, r), (Wild, Wild) | (Rest, Rest) => true, diff --git a/src/tools/clippy/clippy_utils/src/hir_utils.rs b/src/tools/clippy/clippy_utils/src/hir_utils.rs index 9938e64d242..b813cd361ed 100644 --- a/src/tools/clippy/clippy_utils/src/hir_utils.rs +++ b/src/tools/clippy/clippy_utils/src/hir_utils.rs @@ -1124,6 +1124,7 @@ impl<'a, 'tcx> SpanlessHash<'a, 'tcx> { pub fn hash_pat(&mut self, pat: &Pat<'_>) { std::mem::discriminant(&pat.kind).hash(&mut self.s); match pat.kind { + PatKind::Missing => unreachable!(), PatKind::Binding(BindingMode(by_ref, mutability), _, _, pat) => { std::mem::discriminant(&by_ref).hash(&mut self.s); std::mem::discriminant(&mutability).hash(&mut self.s); diff --git a/src/tools/clippy/clippy_utils/src/lib.rs b/src/tools/clippy/clippy_utils/src/lib.rs index 6d2c2a2d692..bcb0f8349e2 100644 --- a/src/tools/clippy/clippy_utils/src/lib.rs +++ b/src/tools/clippy/clippy_utils/src/lib.rs @@ -1858,6 +1858,7 @@ pub fn is_refutable(cx: &LateContext<'_>, pat: &Pat<'_>) -> bool { } match pat.kind { + PatKind::Missing => unreachable!(), PatKind::Wild | PatKind::Never => false, // If `!` typechecked then the type is empty, so not refutable. PatKind::Binding(_, _, _, pat) => pat.is_some_and(|pat| is_refutable(cx, pat)), PatKind::Box(pat) | PatKind::Ref(pat, _) => is_refutable(cx, pat), diff --git a/src/tools/rust-analyzer/Cargo.lock b/src/tools/rust-analyzer/Cargo.lock index 1e1d68f7782..745a8097c8a 100644 --- a/src/tools/rust-analyzer/Cargo.lock +++ b/src/tools/rust-analyzer/Cargo.lock @@ -1270,6 +1270,7 @@ dependencies = [ "edition", "expect-test", "ra-ap-rustc_lexer", + "rustc-literal-escaper", "stdx", "tracing", ] @@ -1744,6 +1745,12 @@ source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "583034fd73374156e66797ed8e5b0d5690409c9226b22d87cb7f19821c05d152" [[package]] +name = "rustc-literal-escaper" +version = "0.0.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0041b6238913c41fe704213a4a9329e2f685a156d1781998128b4149c230ad04" + +[[package]] name = "rustc-stable-hash" version = "0.1.1" source = "registry+https://github.com/rust-lang/crates.io-index" @@ -1978,10 +1985,10 @@ dependencies = [ "indexmap", "itertools", "parser", - "ra-ap-rustc_lexer", "rayon", "rowan", "rustc-hash 2.0.0", + "rustc-literal-escaper", "rustc_apfloat", "smol_str", "stdx", diff --git a/src/tools/rust-analyzer/Cargo.toml b/src/tools/rust-analyzer/Cargo.toml index ce2d66000e3..e2219139765 100644 --- a/src/tools/rust-analyzer/Cargo.toml +++ b/src/tools/rust-analyzer/Cargo.toml @@ -136,6 +136,7 @@ pulldown-cmark-to-cmark = "10.0.4" pulldown-cmark = { version = "0.9.0", default-features = false } rayon = "1.8.0" rustc-hash = "2.0.0" +rustc-literal-escaper = "0.0.2" semver = "1.0.14" serde = { version = "1.0.192" } serde_derive = { version = "1.0.192" } diff --git a/src/tools/rust-analyzer/crates/parser/Cargo.toml b/src/tools/rust-analyzer/crates/parser/Cargo.toml index a36a39dbee6..114a66add63 100644 --- a/src/tools/rust-analyzer/crates/parser/Cargo.toml +++ b/src/tools/rust-analyzer/crates/parser/Cargo.toml @@ -14,6 +14,7 @@ rust-version.workspace = true [dependencies] drop_bomb = "0.1.5" ra-ap-rustc_lexer.workspace = true +rustc-literal-escaper.workspace = true tracing = { workspace = true, optional = true } edition.workspace = true diff --git a/src/tools/rust-analyzer/crates/parser/src/lexed_str.rs b/src/tools/rust-analyzer/crates/parser/src/lexed_str.rs index c97596d5097..b0bbc2fa5ff 100644 --- a/src/tools/rust-analyzer/crates/parser/src/lexed_str.rs +++ b/src/tools/rust-analyzer/crates/parser/src/lexed_str.rs @@ -10,7 +10,7 @@ use std::ops; -use rustc_lexer::unescape::{EscapeError, Mode}; +use rustc_literal_escaper::{EscapeError, Mode, unescape_byte, unescape_char, unescape_mixed, unescape_unicode}; use crate::{ Edition, @@ -282,7 +282,7 @@ impl<'a> Converter<'a> { let text = &self.res.text[self.offset + 1..][..len - 1]; let i = text.rfind('\'').unwrap(); let text = &text[..i]; - if let Err(e) = rustc_lexer::unescape::unescape_char(text) { + if let Err(e) = unescape_char(text) { err = error_to_diagnostic_message(e, Mode::Char); } } @@ -295,7 +295,7 @@ impl<'a> Converter<'a> { let text = &self.res.text[self.offset + 2..][..len - 2]; let i = text.rfind('\'').unwrap(); let text = &text[..i]; - if let Err(e) = rustc_lexer::unescape::unescape_byte(text) { + if let Err(e) = unescape_byte(text) { err = error_to_diagnostic_message(e, Mode::Byte); } } @@ -402,14 +402,14 @@ fn unescape_string_error_message(text: &str, mode: Mode) -> &'static str { let mut error_message = ""; match mode { Mode::CStr => { - rustc_lexer::unescape::unescape_mixed(text, mode, &mut |_, res| { + unescape_mixed(text, mode, &mut |_, res| { if let Err(e) = res { error_message = error_to_diagnostic_message(e, mode); } }); } Mode::ByteStr | Mode::Str => { - rustc_lexer::unescape::unescape_unicode(text, mode, &mut |_, res| { + unescape_unicode(text, mode, &mut |_, res| { if let Err(e) = res { error_message = error_to_diagnostic_message(e, mode); } diff --git a/src/tools/rust-analyzer/crates/syntax/Cargo.toml b/src/tools/rust-analyzer/crates/syntax/Cargo.toml index 3fe6e01dc3c..6b356398204 100644 --- a/src/tools/rust-analyzer/crates/syntax/Cargo.toml +++ b/src/tools/rust-analyzer/crates/syntax/Cargo.toml @@ -17,13 +17,12 @@ either.workspace = true itertools.workspace = true rowan = "=0.15.15" rustc-hash.workspace = true +rustc-literal-escaper.workspace = true indexmap.workspace = true smol_str.workspace = true triomphe.workspace = true tracing.workspace = true -ra-ap-rustc_lexer.workspace = true - parser.workspace = true stdx.workspace = true diff --git a/src/tools/rust-analyzer/crates/syntax/src/ast/token_ext.rs b/src/tools/rust-analyzer/crates/syntax/src/ast/token_ext.rs index df851ab5b25..08bffb9e3aa 100644 --- a/src/tools/rust-analyzer/crates/syntax/src/ast/token_ext.rs +++ b/src/tools/rust-analyzer/crates/syntax/src/ast/token_ext.rs @@ -2,7 +2,7 @@ use std::{borrow::Cow, num::ParseIntError}; -use rustc_lexer::unescape::{ +use rustc_literal_escaper::{ unescape_byte, unescape_char, unescape_mixed, unescape_unicode, EscapeError, MixedUnit, Mode, }; use stdx::always; diff --git a/src/tools/rust-analyzer/crates/syntax/src/lib.rs b/src/tools/rust-analyzer/crates/syntax/src/lib.rs index c9e9f468dca..21f1ea5f913 100644 --- a/src/tools/rust-analyzer/crates/syntax/src/lib.rs +++ b/src/tools/rust-analyzer/crates/syntax/src/lib.rs @@ -19,13 +19,6 @@ //! [RFC]: <https://github.com/rust-lang/rfcs/pull/2256> //! [Swift]: <https://github.com/apple/swift/blob/13d593df6f359d0cb2fc81cfaac273297c539455/lib/Syntax/README.md> -#![cfg_attr(feature = "in-rust-tree", feature(rustc_private))] - -#[cfg(not(feature = "in-rust-tree"))] -extern crate ra_ap_rustc_lexer as rustc_lexer; -#[cfg(feature = "in-rust-tree")] -extern crate rustc_lexer; - mod parsing; mod ptr; mod syntax_error; @@ -64,7 +57,7 @@ pub use rowan::{ api::Preorder, Direction, GreenNode, NodeOrToken, SyntaxText, TextRange, TextSize, TokenAtOffset, WalkEvent, }; -pub use rustc_lexer::unescape; +pub use rustc_literal_escaper as unescape; pub use smol_str::{format_smolstr, SmolStr, SmolStrBuilder, ToSmolStr}; /// `Parse` is the result of the parsing: a syntax tree and a collection of diff --git a/src/tools/rust-analyzer/crates/syntax/src/validation.rs b/src/tools/rust-analyzer/crates/syntax/src/validation.rs index 85eefac734b..71c5f9a946d 100644 --- a/src/tools/rust-analyzer/crates/syntax/src/validation.rs +++ b/src/tools/rust-analyzer/crates/syntax/src/validation.rs @@ -5,7 +5,7 @@ mod block; use rowan::Direction; -use rustc_lexer::unescape::{self, unescape_mixed, unescape_unicode, Mode}; +use rustc_literal_escaper::{unescape_mixed, unescape_unicode, EscapeError, Mode}; use crate::{ algo, @@ -44,8 +44,8 @@ pub(crate) fn validate(root: &SyntaxNode, errors: &mut Vec<SyntaxError>) { } } -fn rustc_unescape_error_to_string(err: unescape::EscapeError) -> (&'static str, bool) { - use unescape::EscapeError as EE; +fn rustc_unescape_error_to_string(err: EscapeError) -> (&'static str, bool) { + use rustc_literal_escaper::EscapeError as EE; #[rustfmt::skip] let err_message = match err { @@ -127,7 +127,7 @@ fn validate_literal(literal: ast::Literal, acc: &mut Vec<SyntaxError>) { let text = token.text(); // FIXME: lift this lambda refactor to `fn` (https://github.com/rust-lang/rust-analyzer/pull/2834#discussion_r366199205) - let mut push_err = |prefix_len, off, err: unescape::EscapeError| { + let mut push_err = |prefix_len, off, err: EscapeError| { let off = token.text_range().start() + TextSize::try_from(off + prefix_len).unwrap(); let (message, is_err) = rustc_unescape_error_to_string(err); // FIXME: Emit lexer warnings diff --git a/src/tools/rustbook/Cargo.lock b/src/tools/rustbook/Cargo.lock index 08011e1d427..09bbbd61de5 100644 --- a/src/tools/rustbook/Cargo.lock +++ b/src/tools/rustbook/Cargo.lock @@ -156,9 +156,9 @@ checksum = "1628fb46dfa0b37568d12e5edd512553eccf6a22a78e8bde00bb4aed84d5bdbf" [[package]] name = "cc" -version = "1.2.17" +version = "1.2.18" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "1fcb57c740ae1daf453ae85f16e37396f672b039e00d9d866e07ddb24e328e3a" +checksum = "525046617d8376e3db1deffb079e91cef90a89fc3ca5c185bbf8c9ecdd15cd5c" dependencies = [ "shlex", ] @@ -185,9 +185,9 @@ dependencies = [ [[package]] name = "clap" -version = "4.5.32" +version = "4.5.35" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "6088f3ae8c3608d19260cd7445411865a485688711b78b5be70d78cd96136f83" +checksum = "d8aa86934b44c19c50f87cc2790e19f54f7a67aedb64101c2e1a2e5ecfb73944" dependencies = [ "clap_builder", "clap_derive", @@ -195,9 +195,9 @@ dependencies = [ [[package]] name = "clap_builder" -version = "4.5.32" +version = "4.5.35" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "22a7ef7f676155edfb82daa97f99441f3ebf4a58d5e32f295a56259f1b6facc8" +checksum = "2414dbb2dd0695280da6ea9261e327479e9d37b0630f6b53ba2a11c60c679fd9" dependencies = [ "anstream", "anstyle", @@ -275,9 +275,9 @@ dependencies = [ [[package]] name = "darling" -version = "0.20.10" +version = "0.20.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "6f63b86c8a8826a49b8c21f08a2d07338eec8d900540f8630dc76284be802989" +checksum = "fc7f46116c46ff9ab3eb1597a45688b6715c6e628b5c133e288e709a29bcb4ee" dependencies = [ "darling_core", "darling_macro", @@ -285,9 +285,9 @@ dependencies = [ [[package]] name = "darling_core" -version = "0.20.10" +version = "0.20.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "95133861a8032aaea082871032f5815eb9e98cef03fa916ab4500513994df9e5" +checksum = "0d00b9596d185e565c2207a0b01f8bd1a135483d02d9b7b0a54b11da8d53412e" dependencies = [ "fnv", "ident_case", @@ -299,9 +299,9 @@ dependencies = [ [[package]] name = "darling_macro" -version = "0.20.10" +version = "0.20.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d336a2a514f6ccccaa3e09b02d41d35330c07ddf03a62165fcec10bb561c7806" +checksum = "fc34b93ccb385b40dc71c6fceac4b2ad23662c7eeb248cf10d529b7e055b6ead" dependencies = [ "darling_core", "quote", @@ -413,9 +413,9 @@ dependencies = [ [[package]] name = "env_logger" -version = "0.11.7" +version = "0.11.8" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "c3716d7a920fb4fac5d84e9d4bce8ceb321e9414b4409da61b07b75c1e3d0697" +checksum = "13c863f0904021b108aa8b2f55046443e6b1ebde8fd4a15c399893aae4fa069f" dependencies = [ "anstream", "anstyle", @@ -432,9 +432,9 @@ checksum = "877a4ace8713b0bcf2a4e7eec82529c029f1d0619886d18145fea96c3ffe5c0f" [[package]] name = "errno" -version = "0.3.10" +version = "0.3.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "33d852cb9b869c2a9b3df2f71a3074817f01e1844f839a144f5fcef059a4eb5d" +checksum = "976dd42dc7e85965fe702eb8164f21f450704bdde31faefd6471dba214cb594e" dependencies = [ "libc", "windows-sys", @@ -455,9 +455,9 @@ checksum = "37909eebbb50d72f9059c3b6d82c0463f2ff062c9e95845c43a6c9c0355411be" [[package]] name = "flate2" -version = "1.1.0" +version = "1.1.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "11faaf5a5236997af9848be0bef4db95824b1d534ebc64d0f0c6cf3e67bd38dc" +checksum = "7ced92e76e966ca2fd84c8f7aa01a4aea65b0eb6648d72f7c8f3e2764a67fece" dependencies = [ "crc32fast", "miniz_oxide", @@ -584,14 +584,15 @@ dependencies = [ [[package]] name = "iana-time-zone" -version = "0.1.61" +version = "0.1.63" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "235e081f3925a06703c2d0117ea8b91f042756fd6e7a6e5d901e8ca1a996b220" +checksum = "b0c919e5debc312ad217002b8048a17b7d83f80703865bbfcfebb0458b0b27d8" dependencies = [ "android_system_properties", "core-foundation-sys", "iana-time-zone-haiku", "js-sys", + "log", "wasm-bindgen", "windows-core", ] @@ -646,9 +647,9 @@ dependencies = [ [[package]] name = "icu_locid_transform_data" -version = "1.5.0" +version = "1.5.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "fdc8ff3388f852bede6b579ad4e978ab004f139284d7b28715f773507b946f6e" +checksum = "7515e6d781098bf9f7205ab3fc7e9709d34554ae0b21ddbcb5febfa4bc7df11d" [[package]] name = "icu_normalizer" @@ -670,9 +671,9 @@ dependencies = [ [[package]] name = "icu_normalizer_data" -version = "1.5.0" +version = "1.5.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "f8cafbf7aa791e9b22bec55a167906f9e1215fd475cd22adfcf660e03e989516" +checksum = "c5e8338228bdc8ab83303f16b797e177953730f601a96c25d10cb3ab0daa0cb7" [[package]] name = "icu_properties" @@ -691,9 +692,9 @@ dependencies = [ [[package]] name = "icu_properties_data" -version = "1.5.0" +version = "1.5.1" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "67a8effbc3dd3e4ba1afa8ad918d5684b8868b3b26500753effea8d2eed19569" +checksum = "85fb8799753b75aee8d2a21d7c14d9f38921b54b3dbda10f5a3c7a7b82dba5e2" [[package]] name = "icu_provider" @@ -752,9 +753,9 @@ dependencies = [ [[package]] name = "indexmap" -version = "2.8.0" +version = "2.9.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "3954d50fe15b02142bf25d3b8bdadb634ec3948f103d04ffe3031bc8fe9d7058" +checksum = "cea70ddb795996207ad57735b50c5982d8844f38ba9ee5f1aedcfb708a2aa11e" dependencies = [ "equivalent", "hashbrown", @@ -861,9 +862,9 @@ dependencies = [ [[package]] name = "log" -version = "0.4.26" +version = "0.4.27" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "30bde2b3dc3671ae49d8e2e9f044c7c005836e7a023ee57cffa25ab82764bb9e" +checksum = "13dc2df351e3202783a1fe0d44375f7295ffb4049267b0f3018346dc122a1d94" [[package]] name = "mac" @@ -980,9 +981,9 @@ checksum = "78ca9ab1a0babb1e7d5695e3530886289c18cf2f87ec19a575a0abdce112e3a3" [[package]] name = "miniz_oxide" -version = "0.8.5" +version = "0.8.7" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8e3e04debbb59698c15bacbb6d93584a8c0ca9cc3213cb423d31f760d8843ce5" +checksum = "ff70ce3e48ae43fa075863cef62e8b43b71a4f2382229920e0df362592919430" dependencies = [ "adler2", ] @@ -1028,9 +1029,9 @@ dependencies = [ [[package]] name = "once_cell" -version = "1.21.1" +version = "1.21.3" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "d75b0bedcc4fe52caa0e03d9f1151a323e4aa5e2d78ba3580400cd3c9e2bc4bc" +checksum = "42f5e15c9953c5e4ccceeb2e7382a716482c34515315f7b03532b8b4e8393d2d" [[package]] name = "onig" @@ -1103,9 +1104,9 @@ checksum = "e3148f5046208a5d56bcfc03053e3ca6334e51da8dfb19b6cdc8b306fae3283e" [[package]] name = "pest" -version = "2.7.15" +version = "2.8.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "8b7cafe60d6cf8e62e1b9b2ea516a089c008945bb5a275416789e7db0bc199dc" +checksum = "198db74531d58c70a361c42201efde7e2591e976d518caf7662a47dc5720e7b6" dependencies = [ "memchr", "thiserror 2.0.12", @@ -1114,9 +1115,9 @@ dependencies = [ [[package]] name = "pest_derive" -version = "2.7.15" +version = "2.8.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "816518421cfc6887a0d62bf441b6ffb4536fcc926395a69e1a85852d4363f57e" +checksum = "d725d9cfd79e87dccc9341a2ef39d1b6f6353d68c4b33c177febbe1a402c97c5" dependencies = [ "pest", "pest_generator", @@ -1124,9 +1125,9 @@ dependencies = [ [[package]] name = "pest_generator" -version = "2.7.15" +version = "2.8.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7d1396fd3a870fc7838768d171b4616d5c91f6cc25e377b673d714567d99377b" +checksum = "db7d01726be8ab66ab32f9df467ae8b1148906685bbe75c82d1e65d7f5b3f841" dependencies = [ "pest", "pest_meta", @@ -1137,9 +1138,9 @@ dependencies = [ [[package]] name = "pest_meta" -version = "2.7.15" +version = "2.8.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e1e58089ea25d717bfd31fb534e4f3afcc2cc569c70de3e239778991ea3b7dea" +checksum = "7f9f832470494906d1fca5329f8ab5791cc60beb230c74815dff541cbd2b5ca0" dependencies = [ "once_cell", "pest", @@ -1316,9 +1317,9 @@ checksum = "ec0be4795e2f6a28069bec0b5ff3e2ac9bafc99e6a9a7dc3547996c5c816922c" [[package]] name = "redox_syscall" -version = "0.5.10" +version = "0.5.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "0b8c0c260b63a8219631167be35e6a988e9554dbd323f8bd08439c8ed1302bd1" +checksum = "d2f103c6d277498fbceb16e84d317e2a400f160f46904d5f5410848c829511a3" dependencies = [ "bitflags 2.9.0", ] @@ -1366,9 +1367,9 @@ dependencies = [ [[package]] name = "rustix" -version = "1.0.3" +version = "1.0.5" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e56a18552996ac8d29ecc3b190b4fdbb2d91ca4ec396de7bbffaf43f3d637e96" +checksum = "d97817398dd4bb2e6da002002db259209759911da105da92bec29ccb12cf58bf" dependencies = [ "bitflags 2.9.0", "errno", @@ -1476,9 +1477,9 @@ checksum = "56199f7ddabf13fe5074ce809e7d3f42b42ae711800501b5b16ea82ad029c39d" [[package]] name = "smallvec" -version = "1.14.0" +version = "1.15.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "7fcf8323ef1faaee30a44a340193b1ac6814fd9b7b4e88e9d4519a3e4abe1cfd" +checksum = "8917285742e9f3e1683f0a9c4e6b57960b7314d0b08d30d1ecd426713ee2eee9" [[package]] name = "stable_deref_trait" @@ -1488,9 +1489,9 @@ checksum = "a8f112729512f8e442d81f95a8a7ddf2b7c6b8a1a6f509a95864142b30cab2d3" [[package]] name = "string_cache" -version = "0.8.8" +version = "0.8.9" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "938d512196766101d333398efde81bc1f37b00cb42c2f8350e5df639f040bbbe" +checksum = "bf776ba3fa74f83bf4b63c3dcbbf82173db2632ed8452cb2d891d33f459de70f" dependencies = [ "new_debug_unreachable", "parking_lot", @@ -1879,11 +1880,37 @@ checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f" [[package]] name = "windows-core" -version = "0.52.0" +version = "0.61.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "33ab640c8d7e35bf8ba19b884ba838ceb4fba93a4e8c65a9059d08afcfc683d9" +checksum = "4763c1de310c86d75a878046489e2e5ba02c649d185f21c67d4cf8a56d098980" dependencies = [ - "windows-targets", + "windows-implement", + "windows-interface", + "windows-link", + "windows-result", + "windows-strings", +] + +[[package]] +name = "windows-implement" +version = "0.60.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a47fddd13af08290e67f4acabf4b459f647552718f683a7b415d290ac744a836" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "windows-interface" +version = "0.59.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bd9211b69f8dcdfa817bfd14bf1c97c9188afa36f4750130fcdf3f400eca9fa8" +dependencies = [ + "proc-macro2", + "quote", + "syn", ] [[package]] @@ -1893,6 +1920,24 @@ source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "76840935b766e1b0a05c0066835fb9ec80071d4c09a16f6bd5f7e655e3c14c38" [[package]] +name = "windows-result" +version = "0.3.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c64fd11a4fd95df68efcfee5f44a294fe71b8bc6a91993e2791938abcc712252" +dependencies = [ + "windows-link", +] + +[[package]] +name = "windows-strings" +version = "0.4.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7a2ba9642430ee452d5a7aa78d72907ebe8cfda358e8cb7918a2050581322f97" +dependencies = [ + "windows-link", +] + +[[package]] name = "windows-sys" version = "0.59.0" source = "registry+https://github.com/rust-lang/crates.io-index" diff --git a/src/tools/rustfmt/src/items.rs b/src/tools/rustfmt/src/items.rs index 322af97d9dc..5c3be769f9e 100644 --- a/src/tools/rustfmt/src/items.rs +++ b/src/tools/rustfmt/src/items.rs @@ -2442,11 +2442,7 @@ pub(crate) fn span_hi_for_param(context: &RewriteContext<'_>, param: &ast::Param } pub(crate) fn is_named_param(param: &ast::Param) -> bool { - if let ast::PatKind::Ident(_, ident, _) = param.pat.kind { - ident.name != symbol::kw::Empty - } else { - true - } + !matches!(param.pat.kind, ast::PatKind::Missing) } #[derive(Copy, Clone, Debug, PartialEq, Eq)] diff --git a/src/tools/rustfmt/src/patterns.rs b/src/tools/rustfmt/src/patterns.rs index 8dc94574503..cb3879f4be8 100644 --- a/src/tools/rustfmt/src/patterns.rs +++ b/src/tools/rustfmt/src/patterns.rs @@ -42,6 +42,7 @@ pub(crate) fn is_short_pattern( fn is_short_pattern_inner(context: &RewriteContext<'_>, pat: &ast::Pat) -> bool { match &pat.kind { + ast::PatKind::Missing => unreachable!(), ast::PatKind::Rest | ast::PatKind::Never | ast::PatKind::Wild | ast::PatKind::Err(_) => { true } @@ -100,6 +101,7 @@ impl Rewrite for Pat { fn rewrite_result(&self, context: &RewriteContext<'_>, shape: Shape) -> RewriteResult { match self.kind { + PatKind::Missing => unreachable!(), PatKind::Or(ref pats) => { let pat_strs = pats .iter() diff --git a/src/tools/tidy/src/deps.rs b/src/tools/tidy/src/deps.rs index 8f761d349cc..682ab4875a1 100644 --- a/src/tools/tidy/src/deps.rs +++ b/src/tools/tidy/src/deps.rs @@ -361,6 +361,7 @@ const PERMITTED_RUSTC_DEPENDENCIES: &[&str] = &[ "regex-syntax", "rustc-demangle", "rustc-hash", + "rustc-literal-escaper", "rustc-rayon", "rustc-rayon-core", "rustc-stable-hash", @@ -486,6 +487,7 @@ const PERMITTED_STDLIB_DEPENDENCIES: &[&str] = &[ "rand_core", "rand_xorshift", "rustc-demangle", + "rustc-literal-escaper", "shlex", "syn", "unicode-ident", diff --git a/tests/pretty/autodiff_forward.pp b/tests/pretty/autodiff_forward.pp index 4b2fb6166ff..713b8f541ae 100644 --- a/tests/pretty/autodiff_forward.pp +++ b/tests/pretty/autodiff_forward.pp @@ -29,6 +29,8 @@ pub fn f1(x: &[f64], y: f64) -> f64 { // Make sure, that we add the None for the default return. + // We want to make sure that we can use the macro for functions defined inside of functions + ::core::panicking::panic("not implemented") } #[rustc_autodiff(Forward, 1, Dual, Const, Dual)] @@ -158,4 +160,25 @@ fn f8_1(x: &f32, bx_0: &f32) -> f32 { ::core::hint::black_box((bx_0,)); ::core::hint::black_box(<f32>::default()) } +pub fn f9() { + #[rustc_autodiff] + #[inline(never)] + fn inner(x: f32) -> f32 { x * x } + #[rustc_autodiff(Forward, 1, Dual, Dual)] + #[inline(never)] + fn d_inner_2(x: f32, bx_0: f32) -> (f32, f32) { + unsafe { asm!("NOP", options(pure, nomem)); }; + ::core::hint::black_box(inner(x)); + ::core::hint::black_box((bx_0,)); + ::core::hint::black_box(<(f32, f32)>::default()) + } + #[rustc_autodiff(Forward, 1, Dual, DualOnly)] + #[inline(never)] + fn d_inner_1(x: f32, bx_0: f32) -> f32 { + unsafe { asm!("NOP", options(pure, nomem)); }; + ::core::hint::black_box(inner(x)); + ::core::hint::black_box((bx_0,)); + ::core::hint::black_box(<f32>::default()) + } +} fn main() {} diff --git a/tests/pretty/autodiff_forward.rs b/tests/pretty/autodiff_forward.rs index a765738c2a8..5a0660a08e5 100644 --- a/tests/pretty/autodiff_forward.rs +++ b/tests/pretty/autodiff_forward.rs @@ -54,4 +54,13 @@ fn f8(x: &f32) -> f32 { unimplemented!() } +// We want to make sure that we can use the macro for functions defined inside of functions +pub fn f9() { + #[autodiff(d_inner_1, Forward, Dual, DualOnly)] + #[autodiff(d_inner_2, Forward, Dual, Dual)] + fn inner(x: f32) -> f32 { + x * x + } +} + fn main() {} diff --git a/tests/ui/attributes/invalid-reprs.rs b/tests/ui/attributes/invalid-reprs.rs new file mode 100644 index 00000000000..95ed14b5491 --- /dev/null +++ b/tests/ui/attributes/invalid-reprs.rs @@ -0,0 +1,6 @@ +fn main() { + let y = #[repr(uwu(4))] + //~^ ERROR attributes on expressions are experimental + //~| ERROR unrecognized representation hint + (&id(5)); //~ ERROR: cannot find function `id` in this scope +} diff --git a/tests/ui/attributes/invalid-reprs.stderr b/tests/ui/attributes/invalid-reprs.stderr new file mode 100644 index 00000000000..415b969b244 --- /dev/null +++ b/tests/ui/attributes/invalid-reprs.stderr @@ -0,0 +1,33 @@ +error[E0658]: attributes on expressions are experimental + --> $DIR/invalid-reprs.rs:2:13 + | +LL | let y = #[repr(uwu(4))] + | ^^^^^^^^^^^^^^^ + | + = note: see issue #15701 <https://github.com/rust-lang/rust/issues/15701> for more information + = help: add `#![feature(stmt_expr_attributes)]` to the crate attributes to enable + = note: this compiler was built on YYYY-MM-DD; consider upgrading it if it is out of date + +error[E0425]: cannot find function `id` in this scope + --> $DIR/invalid-reprs.rs:5:7 + | +LL | (&id(5)); + | ^^ not found in this scope + | +help: consider importing this function + | +LL + use std::process::id; + | + +error[E0552]: unrecognized representation hint + --> $DIR/invalid-reprs.rs:2:20 + | +LL | let y = #[repr(uwu(4))] + | ^^^^^^ + | + = help: valid reprs are `Rust` (default), `C`, `align`, `packed`, `transparent`, `simd`, `i8`, `u8`, `i16`, `u16`, `i32`, `u32`, `i64`, `u64`, `i128`, `u128`, `isize`, `usize` + +error: aborting due to 3 previous errors + +Some errors have detailed explanations: E0425, E0552, E0658. +For more information about an error, try `rustc --explain E0425`. diff --git a/tests/ui/macros/stringify.rs b/tests/ui/macros/stringify.rs index 40033f546d3..3490d3efc59 100644 --- a/tests/ui/macros/stringify.rs +++ b/tests/ui/macros/stringify.rs @@ -515,6 +515,8 @@ fn test_meta() { #[test] fn test_pat() { + // PatKind::Missing: untestable in isolation. + // PatKind::Wild c1!(pat, [ _ ], "_"); diff --git a/tests/ui/unpretty/expanded-exhaustive.rs b/tests/ui/unpretty/expanded-exhaustive.rs index 31af323ecda..4d1f12e3490 100644 --- a/tests/ui/unpretty/expanded-exhaustive.rs +++ b/tests/ui/unpretty/expanded-exhaustive.rs @@ -574,6 +574,11 @@ mod items { } mod patterns { + /// PatKind::Missing + fn pat_missing() { + let _: fn(u32, T, &str); + } + /// PatKind::Wild fn pat_wild() { let _; diff --git a/tests/ui/unpretty/expanded-exhaustive.stdout b/tests/ui/unpretty/expanded-exhaustive.stdout index 19ae66f7a07..d8da941a340 100644 --- a/tests/ui/unpretty/expanded-exhaustive.stdout +++ b/tests/ui/unpretty/expanded-exhaustive.stdout @@ -363,6 +363,7 @@ mod expressions { + { builtin # offset_of(T, field) }; } /// ExprKind::MacCall @@ -517,6 +518,8 @@ mod items { } } mod patterns { + /// PatKind::Missing + fn pat_missing() { let _: fn(u32, T, &str); } /// PatKind::Wild fn pat_wild() { let _; } /// PatKind::Ident |