diff options
Diffstat (limited to 'compiler/rustc_ast/src')
| -rw-r--r-- | compiler/rustc_ast/src/ast.rs | 2 | ||||
| -rw-r--r-- | compiler/rustc_ast/src/lib.rs | 2 | ||||
| -rw-r--r-- | compiler/rustc_ast/src/mut_visit.rs | 140 | ||||
| -rw-r--r-- | compiler/rustc_ast/src/tokenstream.rs | 383 | ||||
| -rw-r--r-- | compiler/rustc_ast/src/visit.rs | 1 |
5 files changed, 388 insertions, 140 deletions
diff --git a/compiler/rustc_ast/src/ast.rs b/compiler/rustc_ast/src/ast.rs index 8986430141b..9d216ef3dd8 100644 --- a/compiler/rustc_ast/src/ast.rs +++ b/compiler/rustc_ast/src/ast.rs @@ -2469,6 +2469,8 @@ pub enum TyPatKind { /// A range pattern (e.g., `1...2`, `1..2`, `1..`, `..2`, `1..=2`, `..=2`). Range(Option<P<AnonConst>>, Option<P<AnonConst>>, Spanned<RangeEnd>), + Or(ThinVec<P<TyPat>>), + /// Placeholder for a pattern that wasn't syntactically well formed in some way. Err(ErrorGuaranteed), } diff --git a/compiler/rustc_ast/src/lib.rs b/compiler/rustc_ast/src/lib.rs index 1471262d2d6..e572ec99dab 100644 --- a/compiler/rustc_ast/src/lib.rs +++ b/compiler/rustc_ast/src/lib.rs @@ -12,6 +12,7 @@ test(attr(deny(warnings))) )] #![doc(rust_logo)] +#![feature(array_windows)] #![feature(associated_type_defaults)] #![feature(box_patterns)] #![feature(if_let_guard)] @@ -19,6 +20,7 @@ #![feature(never_type)] #![feature(rustdoc_internals)] #![feature(stmt_expr_attributes)] +#![recursion_limit = "256"] // tidy-alphabetical-end pub mod util { diff --git a/compiler/rustc_ast/src/mut_visit.rs b/compiler/rustc_ast/src/mut_visit.rs index 6aae2e481a5..e49886721e3 100644 --- a/compiler/rustc_ast/src/mut_visit.rs +++ b/compiler/rustc_ast/src/mut_visit.rs @@ -9,7 +9,6 @@ use std::ops::DerefMut; use std::panic; -use std::sync::Arc; use rustc_data_structures::flat_map_in_place::FlatMapInPlace; use rustc_data_structures::stack::ensure_sufficient_stack; @@ -20,7 +19,6 @@ use thin_vec::ThinVec; use crate::ast::*; use crate::ptr::P; -use crate::token::{self, Token}; use crate::tokenstream::*; use crate::visit::{AssocCtxt, BoundKind, FnCtxt}; @@ -48,11 +46,6 @@ pub trait WalkItemKind { } pub trait MutVisitor: Sized { - /// Mutable token visiting only exists for the `macro_rules` token marker and should not be - /// used otherwise. Token visitor would be entirely separate from the regular visitor if - /// the marker didn't have to visit AST fragments in nonterminal tokens. - const VISIT_TOKENS: bool = false; - // Methods in this trait have one of three forms: // // fn visit_t(&mut self, t: &mut T); // common @@ -360,6 +353,8 @@ pub trait MutVisitor: Sized { // Do nothing. } + // Span visiting is no longer used, but we keep it for now, + // in case it's needed for something like #127241. fn visit_span(&mut self, _sp: &mut Span) { // Do nothing. } @@ -473,12 +468,8 @@ fn visit_attr_args<T: MutVisitor>(vis: &mut T, args: &mut AttrArgs) { // No `noop_` prefix because there isn't a corresponding method in `MutVisitor`. fn visit_delim_args<T: MutVisitor>(vis: &mut T, args: &mut DelimArgs) { - let DelimArgs { dspan, delim: _, tokens } = args; - visit_tts(vis, tokens); - visit_delim_span(vis, dspan); -} - -pub fn visit_delim_span<T: MutVisitor>(vis: &mut T, DelimSpan { open, close }: &mut DelimSpan) { + let DelimArgs { dspan, delim: _, tokens: _ } = args; + let DelimSpan { open, close } = dspan; vis.visit_span(open); vis.visit_span(close); } @@ -552,7 +543,7 @@ fn walk_assoc_item_constraint<T: MutVisitor>( } pub fn walk_ty<T: MutVisitor>(vis: &mut T, ty: &mut P<Ty>) { - let Ty { id, kind, span, tokens } = ty.deref_mut(); + let Ty { id, kind, span, tokens: _ } = ty.deref_mut(); vis.visit_id(id); match kind { TyKind::Err(_guar) => {} @@ -600,21 +591,20 @@ pub fn walk_ty<T: MutVisitor>(vis: &mut T, ty: &mut P<Ty>) { } TyKind::MacCall(mac) => vis.visit_mac_call(mac), } - visit_lazy_tts(vis, tokens); vis.visit_span(span); } pub fn walk_ty_pat<T: MutVisitor>(vis: &mut T, ty: &mut P<TyPat>) { - let TyPat { id, kind, span, tokens } = ty.deref_mut(); + let TyPat { id, kind, span, tokens: _ } = ty.deref_mut(); vis.visit_id(id); match kind { TyPatKind::Range(start, end, _include_end) => { visit_opt(start, |c| vis.visit_anon_const(c)); visit_opt(end, |c| vis.visit_anon_const(c)); } + TyPatKind::Or(variants) => visit_thin_vec(variants, |p| vis.visit_ty_pat(p)), TyPatKind::Err(_) => {} } - visit_lazy_tts(vis, tokens); vis.visit_span(span); } @@ -654,11 +644,10 @@ fn walk_path_segment<T: MutVisitor>(vis: &mut T, segment: &mut PathSegment) { visit_opt(args, |args| vis.visit_generic_args(args)); } -fn walk_path<T: MutVisitor>(vis: &mut T, Path { segments, span, tokens }: &mut Path) { +fn walk_path<T: MutVisitor>(vis: &mut T, Path { segments, span, tokens: _ }: &mut Path) { for segment in segments { vis.visit_path_segment(segment); } - visit_lazy_tts(vis, tokens); vis.visit_span(span); } @@ -704,7 +693,7 @@ fn walk_parenthesized_parameter_data<T: MutVisitor>(vis: &mut T, args: &mut Pare } fn walk_local<T: MutVisitor>(vis: &mut T, local: &mut P<Local>) { - let Local { id, super_, pat, ty, kind, span, colon_sp, attrs, tokens } = local.deref_mut(); + let Local { id, super_, pat, ty, kind, span, colon_sp, attrs, tokens: _ } = local.deref_mut(); visit_opt(super_, |sp| vis.visit_span(sp)); vis.visit_id(id); visit_attrs(vis, attrs); @@ -720,7 +709,6 @@ fn walk_local<T: MutVisitor>(vis: &mut T, local: &mut P<Local>) { vis.visit_block(els); } } - visit_lazy_tts(vis, tokens); visit_opt(colon_sp, |sp| vis.visit_span(sp)); vis.visit_span(span); } @@ -729,14 +717,10 @@ fn walk_attribute<T: MutVisitor>(vis: &mut T, attr: &mut Attribute) { let Attribute { kind, id: _, style: _, span } = attr; match kind { AttrKind::Normal(normal) => { - let NormalAttr { - item: AttrItem { unsafety: _, path, args, tokens }, - tokens: attr_tokens, - } = &mut **normal; + let NormalAttr { item: AttrItem { unsafety: _, path, args, tokens: _ }, tokens: _ } = + &mut **normal; vis.visit_path(path); visit_attr_args(vis, args); - visit_lazy_tts(vis, tokens); - visit_lazy_tts(vis, attr_tokens); } AttrKind::DocComment(_kind, _sym) => {} } @@ -786,90 +770,6 @@ pub fn walk_flat_map_param<T: MutVisitor>(vis: &mut T, mut param: Param) -> Smal } // No `noop_` prefix because there isn't a corresponding method in `MutVisitor`. -fn visit_attr_tt<T: MutVisitor>(vis: &mut T, tt: &mut AttrTokenTree) { - match tt { - AttrTokenTree::Token(token, _spacing) => { - visit_token(vis, token); - } - AttrTokenTree::Delimited(dspan, _spacing, _delim, tts) => { - visit_attr_tts(vis, tts); - visit_delim_span(vis, dspan); - } - AttrTokenTree::AttrsTarget(AttrsTarget { attrs, tokens }) => { - visit_attrs(vis, attrs); - visit_lazy_tts_opt_mut(vis, Some(tokens)); - } - } -} - -// No `noop_` prefix because there isn't a corresponding method in `MutVisitor`. -fn visit_tt<T: MutVisitor>(vis: &mut T, tt: &mut TokenTree) { - match tt { - TokenTree::Token(token, _spacing) => { - visit_token(vis, token); - } - TokenTree::Delimited(dspan, _spacing, _delim, tts) => { - visit_tts(vis, tts); - visit_delim_span(vis, dspan); - } - } -} - -// No `noop_` prefix because there isn't a corresponding method in `MutVisitor`. -fn visit_tts<T: MutVisitor>(vis: &mut T, TokenStream(tts): &mut TokenStream) { - if T::VISIT_TOKENS && !tts.is_empty() { - let tts = Arc::make_mut(tts); - visit_vec(tts, |tree| visit_tt(vis, tree)); - } -} - -fn visit_attr_tts<T: MutVisitor>(vis: &mut T, AttrTokenStream(tts): &mut AttrTokenStream) { - if T::VISIT_TOKENS && !tts.is_empty() { - let tts = Arc::make_mut(tts); - visit_vec(tts, |tree| visit_attr_tt(vis, tree)); - } -} - -fn visit_lazy_tts_opt_mut<T: MutVisitor>(vis: &mut T, lazy_tts: Option<&mut LazyAttrTokenStream>) { - if T::VISIT_TOKENS { - if let Some(lazy_tts) = lazy_tts { - let mut tts = lazy_tts.to_attr_token_stream(); - visit_attr_tts(vis, &mut tts); - *lazy_tts = LazyAttrTokenStream::new(tts); - } - } -} - -fn visit_lazy_tts<T: MutVisitor>(vis: &mut T, lazy_tts: &mut Option<LazyAttrTokenStream>) { - visit_lazy_tts_opt_mut(vis, lazy_tts.as_mut()); -} - -/// Applies ident visitor if it's an ident. In practice this is not actually -/// used by specific visitors right now, but there's a test below checking that -/// it works. -// No `noop_` prefix because there isn't a corresponding method in `MutVisitor`. -pub fn visit_token<T: MutVisitor>(vis: &mut T, t: &mut Token) { - let Token { kind, span } = t; - match kind { - token::Ident(name, _is_raw) | token::Lifetime(name, _is_raw) => { - let mut ident = Ident::new(*name, *span); - vis.visit_ident(&mut ident); - *name = ident.name; - *span = ident.span; - return; // Avoid visiting the span for the second time. - } - token::NtIdent(ident, _is_raw) => { - vis.visit_ident(ident); - } - token::NtLifetime(ident, _is_raw) => { - vis.visit_ident(ident); - } - _ => {} - } - vis.visit_span(span); -} - -// No `noop_` prefix because there isn't a corresponding method in `MutVisitor`. fn visit_defaultness<T: MutVisitor>(vis: &mut T, defaultness: &mut Defaultness) { match defaultness { Defaultness::Default(span) => vis.visit_span(span), @@ -1187,10 +1087,9 @@ fn walk_mt<T: MutVisitor>(vis: &mut T, MutTy { ty, mutbl: _ }: &mut MutTy) { } pub fn walk_block<T: MutVisitor>(vis: &mut T, block: &mut P<Block>) { - let Block { id, stmts, rules: _, span, tokens } = block.deref_mut(); + let Block { id, stmts, rules: _, span, tokens: _ } = block.deref_mut(); vis.visit_id(id); stmts.flat_map_in_place(|stmt| vis.flat_map_stmt(stmt)); - visit_lazy_tts(vis, tokens); vis.visit_span(span); } @@ -1471,12 +1370,11 @@ fn walk_item_ctxt<K: WalkItemKind>( item: &mut P<Item<K>>, ctxt: K::Ctxt, ) { - let Item { attrs, id, kind, vis, span, tokens } = item.deref_mut(); + let Item { attrs, id, kind, vis, span, tokens: _ } = item.deref_mut(); visitor.visit_id(id); visit_attrs(visitor, attrs); visitor.visit_vis(vis); kind.walk(*span, *id, vis, ctxt, visitor); - visit_lazy_tts(visitor, tokens); visitor.visit_span(span); } @@ -1550,7 +1448,7 @@ impl WalkItemKind for ForeignItemKind { } pub fn walk_pat<T: MutVisitor>(vis: &mut T, pat: &mut P<Pat>) { - let Pat { id, kind, span, tokens } = pat.deref_mut(); + let Pat { id, kind, span, tokens: _ } = pat.deref_mut(); vis.visit_id(id); match kind { PatKind::Err(_guar) => {} @@ -1592,7 +1490,6 @@ pub fn walk_pat<T: MutVisitor>(vis: &mut T, pat: &mut P<Pat>) { PatKind::Paren(inner) => vis.visit_pat(inner), PatKind::MacCall(mac) => vis.visit_mac_call(mac), } - visit_lazy_tts(vis, tokens); vis.visit_span(span); } @@ -1656,7 +1553,7 @@ fn walk_format_args<T: MutVisitor>(vis: &mut T, fmt: &mut FormatArgs) { vis.visit_span(span); } -pub fn walk_expr<T: MutVisitor>(vis: &mut T, Expr { kind, id, span, attrs, tokens }: &mut Expr) { +pub fn walk_expr<T: MutVisitor>(vis: &mut T, Expr { kind, id, span, attrs, tokens: _ }: &mut Expr) { vis.visit_id(id); visit_attrs(vis, attrs); match kind { @@ -1847,7 +1744,6 @@ pub fn walk_expr<T: MutVisitor>(vis: &mut T, Expr { kind, id, span, attrs, token ExprKind::Err(_guar) => {} ExprKind::Dummy => {} } - visit_lazy_tts(vis, tokens); vis.visit_span(span); } @@ -1889,17 +1785,16 @@ fn walk_flat_map_stmt_kind<T: MutVisitor>(vis: &mut T, kind: StmtKind) -> SmallV StmtKind::Semi(expr) => vis.filter_map_expr(expr).into_iter().map(StmtKind::Semi).collect(), StmtKind::Empty => smallvec![StmtKind::Empty], StmtKind::MacCall(mut mac) => { - let MacCallStmt { mac: mac_, style: _, attrs, tokens } = mac.deref_mut(); + let MacCallStmt { mac: mac_, style: _, attrs, tokens: _ } = mac.deref_mut(); visit_attrs(vis, attrs); vis.visit_mac_call(mac_); - visit_lazy_tts(vis, tokens); smallvec![StmtKind::MacCall(mac)] } } } fn walk_vis<T: MutVisitor>(vis: &mut T, visibility: &mut Visibility) { - let Visibility { kind, span, tokens } = visibility; + let Visibility { kind, span, tokens: _ } = visibility; match kind { VisibilityKind::Public | VisibilityKind::Inherited => {} VisibilityKind::Restricted { path, id, shorthand: _ } => { @@ -1907,7 +1802,6 @@ fn walk_vis<T: MutVisitor>(vis: &mut T, visibility: &mut Visibility) { vis.visit_path(path); } } - visit_lazy_tts(vis, tokens); vis.visit_span(span); } diff --git a/compiler/rustc_ast/src/tokenstream.rs b/compiler/rustc_ast/src/tokenstream.rs index 43d25d18075..636c26bcde0 100644 --- a/compiler/rustc_ast/src/tokenstream.rs +++ b/compiler/rustc_ast/src/tokenstream.rs @@ -14,14 +14,16 @@ //! ownership of the original. use std::borrow::Cow; +use std::ops::Range; use std::sync::Arc; -use std::{cmp, fmt, iter}; +use std::{cmp, fmt, iter, mem}; use rustc_data_structures::stable_hasher::{HashStable, StableHasher}; use rustc_data_structures::sync; use rustc_macros::{Decodable, Encodable, HashStable_Generic}; use rustc_serialize::{Decodable, Encodable}; use rustc_span::{DUMMY_SP, Span, SpanDecoder, SpanEncoder, Symbol, sym}; +use thin_vec::ThinVec; use crate::ast::AttrStyle; use crate::ast_traits::{HasAttrs, HasTokens}; @@ -106,25 +108,30 @@ where } } -pub trait ToAttrTokenStream: sync::DynSend + sync::DynSync { - fn to_attr_token_stream(&self) -> AttrTokenStream; -} - -impl ToAttrTokenStream for AttrTokenStream { - fn to_attr_token_stream(&self) -> AttrTokenStream { - self.clone() - } -} - -/// A lazy version of [`TokenStream`], which defers creation -/// of an actual `TokenStream` until it is needed. -/// `Box` is here only to reduce the structure size. +/// A lazy version of [`AttrTokenStream`], which defers creation of an actual +/// `AttrTokenStream` until it is needed. #[derive(Clone)] -pub struct LazyAttrTokenStream(Arc<Box<dyn ToAttrTokenStream>>); +pub struct LazyAttrTokenStream(Arc<LazyAttrTokenStreamInner>); impl LazyAttrTokenStream { - pub fn new(inner: impl ToAttrTokenStream + 'static) -> LazyAttrTokenStream { - LazyAttrTokenStream(Arc::new(Box::new(inner))) + pub fn new_direct(stream: AttrTokenStream) -> LazyAttrTokenStream { + LazyAttrTokenStream(Arc::new(LazyAttrTokenStreamInner::Direct(stream))) + } + + pub fn new_pending( + start_token: (Token, Spacing), + cursor_snapshot: TokenCursor, + num_calls: u32, + break_last_token: u32, + node_replacements: ThinVec<NodeReplacement>, + ) -> LazyAttrTokenStream { + LazyAttrTokenStream(Arc::new(LazyAttrTokenStreamInner::Pending { + start_token, + cursor_snapshot, + num_calls, + break_last_token, + node_replacements, + })) } pub fn to_attr_token_stream(&self) -> AttrTokenStream { @@ -156,6 +163,184 @@ impl<CTX> HashStable<CTX> for LazyAttrTokenStream { } } +/// A token range within a `Parser`'s full token stream. +#[derive(Clone, Debug)] +pub struct ParserRange(pub Range<u32>); + +/// A token range within an individual AST node's (lazy) token stream, i.e. +/// relative to that node's first token. Distinct from `ParserRange` so the two +/// kinds of range can't be mixed up. +#[derive(Clone, Debug)] +pub struct NodeRange(pub Range<u32>); + +/// Indicates a range of tokens that should be replaced by an `AttrsTarget` +/// (replacement) or be replaced by nothing (deletion). This is used in two +/// places during token collection. +/// +/// 1. Replacement. During the parsing of an AST node that may have a +/// `#[derive]` attribute, when we parse a nested AST node that has `#[cfg]` +/// or `#[cfg_attr]`, we replace the entire inner AST node with +/// `FlatToken::AttrsTarget`. This lets us perform eager cfg-expansion on an +/// `AttrTokenStream`. +/// +/// 2. Deletion. We delete inner attributes from all collected token streams, +/// and instead track them through the `attrs` field on the AST node. This +/// lets us manipulate them similarly to outer attributes. When we create a +/// `TokenStream`, the inner attributes are inserted into the proper place +/// in the token stream. +/// +/// Each replacement starts off in `ParserReplacement` form but is converted to +/// `NodeReplacement` form when it is attached to a single AST node, via +/// `LazyAttrTokenStreamImpl`. +pub type ParserReplacement = (ParserRange, Option<AttrsTarget>); + +/// See the comment on `ParserReplacement`. +pub type NodeReplacement = (NodeRange, Option<AttrsTarget>); + +impl NodeRange { + // Converts a range within a parser's tokens to a range within a + // node's tokens beginning at `start_pos`. + // + // For example, imagine a parser with 50 tokens in its token stream, a + // function that spans `ParserRange(20..40)` and an inner attribute within + // that function that spans `ParserRange(30..35)`. We would find the inner + // attribute's range within the function's tokens by subtracting 20, which + // is the position of the function's start token. This gives + // `NodeRange(10..15)`. + pub fn new(ParserRange(parser_range): ParserRange, start_pos: u32) -> NodeRange { + assert!(!parser_range.is_empty()); + assert!(parser_range.start >= start_pos); + NodeRange((parser_range.start - start_pos)..(parser_range.end - start_pos)) + } +} + +enum LazyAttrTokenStreamInner { + // The token stream has already been produced. + Direct(AttrTokenStream), + + // From a value of this type we can reconstruct the `TokenStream` seen by + // the `f` callback passed to a call to `Parser::collect_tokens`, by + // replaying the getting of the tokens. This saves us producing a + // `TokenStream` if it is never needed, e.g. a captured `macro_rules!` + // argument that is never passed to a proc macro. In practice, token stream + // creation happens rarely compared to calls to `collect_tokens` (see some + // statistics in #78736) so we are doing as little up-front work as + // possible. + // + // This also makes `Parser` very cheap to clone, since there is no + // intermediate collection buffer to clone. + Pending { + start_token: (Token, Spacing), + cursor_snapshot: TokenCursor, + num_calls: u32, + break_last_token: u32, + node_replacements: ThinVec<NodeReplacement>, + }, +} + +impl LazyAttrTokenStreamInner { + fn to_attr_token_stream(&self) -> AttrTokenStream { + match self { + LazyAttrTokenStreamInner::Direct(stream) => stream.clone(), + LazyAttrTokenStreamInner::Pending { + start_token, + cursor_snapshot, + num_calls, + break_last_token, + node_replacements, + } => { + // The token produced by the final call to `{,inlined_}next` was not + // actually consumed by the callback. The combination of chaining the + // initial token and using `take` produces the desired result - we + // produce an empty `TokenStream` if no calls were made, and omit the + // final token otherwise. + let mut cursor_snapshot = cursor_snapshot.clone(); + let tokens = iter::once(FlatToken::Token(*start_token)) + .chain(iter::repeat_with(|| FlatToken::Token(cursor_snapshot.next()))) + .take(*num_calls as usize); + + if node_replacements.is_empty() { + make_attr_token_stream(tokens, *break_last_token) + } else { + let mut tokens: Vec<_> = tokens.collect(); + let mut node_replacements = node_replacements.to_vec(); + node_replacements.sort_by_key(|(range, _)| range.0.start); + + #[cfg(debug_assertions)] + for [(node_range, tokens), (next_node_range, next_tokens)] in + node_replacements.array_windows() + { + assert!( + node_range.0.end <= next_node_range.0.start + || node_range.0.end >= next_node_range.0.end, + "Node ranges should be disjoint or nested: ({:?}, {:?}) ({:?}, {:?})", + node_range, + tokens, + next_node_range, + next_tokens, + ); + } + + // Process the replace ranges, starting from the highest start + // position and working our way back. If have tokens like: + // + // `#[cfg(FALSE)] struct Foo { #[cfg(FALSE)] field: bool }` + // + // Then we will generate replace ranges for both + // the `#[cfg(FALSE)] field: bool` and the entire + // `#[cfg(FALSE)] struct Foo { #[cfg(FALSE)] field: bool }` + // + // By starting processing from the replace range with the greatest + // start position, we ensure that any (outer) replace range which + // encloses another (inner) replace range will fully overwrite the + // inner range's replacement. + for (node_range, target) in node_replacements.into_iter().rev() { + assert!( + !node_range.0.is_empty(), + "Cannot replace an empty node range: {:?}", + node_range.0 + ); + + // Replace the tokens in range with zero or one `FlatToken::AttrsTarget`s, + // plus enough `FlatToken::Empty`s to fill up the rest of the range. This + // keeps the total length of `tokens` constant throughout the replacement + // process, allowing us to do all replacements without adjusting indices. + let target_len = target.is_some() as usize; + tokens.splice( + (node_range.0.start as usize)..(node_range.0.end as usize), + target.into_iter().map(|target| FlatToken::AttrsTarget(target)).chain( + iter::repeat(FlatToken::Empty) + .take(node_range.0.len() - target_len), + ), + ); + } + make_attr_token_stream(tokens.into_iter(), *break_last_token) + } + } + } + } +} + +/// A helper struct used when building an `AttrTokenStream` from +/// a `LazyAttrTokenStream`. Both delimiter and non-delimited tokens +/// are stored as `FlatToken::Token`. A vector of `FlatToken`s +/// is then 'parsed' to build up an `AttrTokenStream` with nested +/// `AttrTokenTree::Delimited` tokens. +#[derive(Debug, Clone)] +enum FlatToken { + /// A token - this holds both delimiter (e.g. '{' and '}') + /// and non-delimiter tokens + Token((Token, Spacing)), + /// Holds the `AttrsTarget` for an AST node. The `AttrsTarget` is inserted + /// directly into the constructed `AttrTokenStream` as an + /// `AttrTokenTree::AttrsTarget`. + AttrsTarget(AttrsTarget), + /// A special 'empty' token that is ignored during the conversion + /// to an `AttrTokenStream`. This is used to simplify the + /// handling of replace ranges. + Empty, +} + /// An `AttrTokenStream` is similar to a `TokenStream`, but with extra /// information about the tokens for attribute targets. This is used /// during expansion to perform early cfg-expansion, and to process attributes @@ -163,6 +348,71 @@ impl<CTX> HashStable<CTX> for LazyAttrTokenStream { #[derive(Clone, Debug, Default, Encodable, Decodable)] pub struct AttrTokenStream(pub Arc<Vec<AttrTokenTree>>); +/// Converts a flattened iterator of tokens (including open and close delimiter tokens) into an +/// `AttrTokenStream`, creating an `AttrTokenTree::Delimited` for each matching pair of open and +/// close delims. +fn make_attr_token_stream( + iter: impl Iterator<Item = FlatToken>, + break_last_token: u32, +) -> AttrTokenStream { + #[derive(Debug)] + struct FrameData { + // This is `None` for the first frame, `Some` for all others. + open_delim_sp: Option<(Delimiter, Span, Spacing)>, + inner: Vec<AttrTokenTree>, + } + // The stack always has at least one element. Storing it separately makes for shorter code. + let mut stack_top = FrameData { open_delim_sp: None, inner: vec![] }; + let mut stack_rest = vec![]; + for flat_token in iter { + match flat_token { + FlatToken::Token((token @ Token { kind, span }, spacing)) => { + if let Some(delim) = kind.open_delim() { + stack_rest.push(mem::replace( + &mut stack_top, + FrameData { open_delim_sp: Some((delim, span, spacing)), inner: vec![] }, + )); + } else if let Some(delim) = kind.close_delim() { + let frame_data = mem::replace(&mut stack_top, stack_rest.pop().unwrap()); + let (open_delim, open_sp, open_spacing) = frame_data.open_delim_sp.unwrap(); + assert!( + open_delim.eq_ignoring_invisible_origin(&delim), + "Mismatched open/close delims: open={open_delim:?} close={span:?}" + ); + let dspan = DelimSpan::from_pair(open_sp, span); + let dspacing = DelimSpacing::new(open_spacing, spacing); + let stream = AttrTokenStream::new(frame_data.inner); + let delimited = AttrTokenTree::Delimited(dspan, dspacing, delim, stream); + stack_top.inner.push(delimited); + } else { + stack_top.inner.push(AttrTokenTree::Token(token, spacing)) + } + } + FlatToken::AttrsTarget(target) => { + stack_top.inner.push(AttrTokenTree::AttrsTarget(target)) + } + FlatToken::Empty => {} + } + } + + if break_last_token > 0 { + let last_token = stack_top.inner.pop().unwrap(); + if let AttrTokenTree::Token(last_token, spacing) = last_token { + let (unglued, _) = last_token.kind.break_two_token_op(break_last_token).unwrap(); + + // Tokens are always ASCII chars, so we can use byte arithmetic here. + let mut first_span = last_token.span.shrink_to_lo(); + first_span = + first_span.with_hi(first_span.lo() + rustc_span::BytePos(break_last_token)); + + stack_top.inner.push(AttrTokenTree::Token(Token::new(unglued, first_span), spacing)); + } else { + panic!("Unexpected last token {last_token:?}") + } + } + AttrTokenStream::new(stack_top.inner) +} + /// Like `TokenTree`, but for `AttrTokenStream`. #[derive(Clone, Debug, Encodable, Decodable)] pub enum AttrTokenTree { @@ -641,6 +891,104 @@ impl<'t> Iterator for TokenStreamIter<'t> { } } +#[derive(Clone, Debug)] +pub struct TokenTreeCursor { + stream: TokenStream, + /// Points to the current token tree in the stream. In `TokenCursor::curr`, + /// this can be any token tree. In `TokenCursor::stack`, this is always a + /// `TokenTree::Delimited`. + index: usize, +} + +impl TokenTreeCursor { + #[inline] + pub fn new(stream: TokenStream) -> Self { + TokenTreeCursor { stream, index: 0 } + } + + #[inline] + pub fn curr(&self) -> Option<&TokenTree> { + self.stream.get(self.index) + } + + pub fn look_ahead(&self, n: usize) -> Option<&TokenTree> { + self.stream.get(self.index + n) + } + + #[inline] + pub fn bump(&mut self) { + self.index += 1; + } +} + +/// A `TokenStream` cursor that produces `Token`s. It's a bit odd that +/// we (a) lex tokens into a nice tree structure (`TokenStream`), and then (b) +/// use this type to emit them as a linear sequence. But a linear sequence is +/// what the parser expects, for the most part. +#[derive(Clone, Debug)] +pub struct TokenCursor { + // Cursor for the current (innermost) token stream. The index within the + // cursor can point to any token tree in the stream (or one past the end). + // The delimiters for this token stream are found in `self.stack.last()`; + // if that is `None` we are in the outermost token stream which never has + // delimiters. + pub curr: TokenTreeCursor, + + // Token streams surrounding the current one. The index within each cursor + // always points to a `TokenTree::Delimited`. + pub stack: Vec<TokenTreeCursor>, +} + +impl TokenCursor { + pub fn next(&mut self) -> (Token, Spacing) { + self.inlined_next() + } + + /// This always-inlined version should only be used on hot code paths. + #[inline(always)] + pub fn inlined_next(&mut self) -> (Token, Spacing) { + loop { + // FIXME: we currently don't return `Delimiter::Invisible` open/close delims. To fix + // #67062 we will need to, whereupon the `delim != Delimiter::Invisible` conditions + // below can be removed. + if let Some(tree) = self.curr.curr() { + match tree { + &TokenTree::Token(token, spacing) => { + debug_assert!(!token.kind.is_delim()); + let res = (token, spacing); + self.curr.bump(); + return res; + } + &TokenTree::Delimited(sp, spacing, delim, ref tts) => { + let trees = TokenTreeCursor::new(tts.clone()); + self.stack.push(mem::replace(&mut self.curr, trees)); + if !delim.skip() { + return (Token::new(delim.as_open_token_kind(), sp.open), spacing.open); + } + // No open delimiter to return; continue on to the next iteration. + } + }; + } else if let Some(parent) = self.stack.pop() { + // We have exhausted this token stream. Move back to its parent token stream. + let Some(&TokenTree::Delimited(span, spacing, delim, _)) = parent.curr() else { + panic!("parent should be Delimited") + }; + self.curr = parent; + self.curr.bump(); // move past the `Delimited` + if !delim.skip() { + return (Token::new(delim.as_close_token_kind(), span.close), spacing.close); + } + // No close delimiter to return; continue on to the next iteration. + } else { + // We have exhausted the outermost token stream. The use of + // `Spacing::Alone` is arbitrary and immaterial, because the + // `Eof` token's spacing is never used. + return (Token::new(token::Eof, DUMMY_SP), Spacing::Alone); + } + } + } +} + #[derive(Debug, Copy, Clone, PartialEq, Encodable, Decodable, HashStable_Generic)] pub struct DelimSpan { pub open: Span, @@ -687,6 +1035,7 @@ mod size_asserts { static_assert_size!(AttrTokenStream, 8); static_assert_size!(AttrTokenTree, 32); static_assert_size!(LazyAttrTokenStream, 8); + static_assert_size!(LazyAttrTokenStreamInner, 88); static_assert_size!(Option<LazyAttrTokenStream>, 8); // must be small, used in many AST nodes static_assert_size!(TokenStream, 8); static_assert_size!(TokenTree, 32); diff --git a/compiler/rustc_ast/src/visit.rs b/compiler/rustc_ast/src/visit.rs index 79193fcec63..69a186c8cf1 100644 --- a/compiler/rustc_ast/src/visit.rs +++ b/compiler/rustc_ast/src/visit.rs @@ -608,6 +608,7 @@ pub fn walk_ty_pat<'a, V: Visitor<'a>>(visitor: &mut V, tp: &'a TyPat) -> V::Res visit_opt!(visitor, visit_anon_const, start); visit_opt!(visitor, visit_anon_const, end); } + TyPatKind::Or(variants) => walk_list!(visitor, visit_ty_pat, variants), TyPatKind::Err(_) => {} } V::Result::output() |