Call compute_locs once per rule.

Currently it's called in `parse_tt` every time a match rule is invoked.
This commit moves it so it's called instead once per match rule, in
`compile_declarative_macro. This is a performance win.

The commit also moves `compute_locs` out of `TtParser`, because there's
no longer any reason for it to be in there.

Author: nnethercote

compiler/rustc_expand/src/mbe/macro_parser.rs

@@ -106,7 +106,7 @@ rustc_data_structures::static_assert_size!(NamedMatchVec, 48);
 ///
 /// This means a matcher can be represented by `&[MatcherLoc]`, and traversal mostly involves
 /// simply incrementing the current matcher position index by one.
-enum MatcherLoc {
+pub(super) enum MatcherLoc {
     Token {
         token: Token,
     },
@@ -138,6 +138,78 @@ enum MatcherLoc {
     Eof,
 }
 
+pub(super) fn compute_locs(sess: &ParseSess, matcher: &[TokenTree]) -> Vec<MatcherLoc> {
+    fn inner(
+        sess: &ParseSess,
+        tts: &[TokenTree],
+        locs: &mut Vec<MatcherLoc>,
+        next_metavar: &mut usize,
+        seq_depth: usize,
+    ) {
+        for tt in tts {
+            match tt {
+                TokenTree::Token(token) => {
+                    locs.push(MatcherLoc::Token { token: token.clone() });
+                }
+                TokenTree::Delimited(_, delimited) => {
+                    locs.push(MatcherLoc::Delimited);
+                    inner(sess, &delimited.all_tts, locs, next_metavar, seq_depth);
+                }
+                TokenTree::Sequence(_, seq) => {
+                    // We can't determine `idx_first_after` and construct the final
+                    // `MatcherLoc::Sequence` until after `inner()` is called and the sequence end
+                    // pieces are processed. So we push a dummy value (`Eof` is cheapest to
+                    // construct) now, and overwrite it with the proper value below.
+                    let dummy = MatcherLoc::Eof;
+                    locs.push(dummy);
+
+                    let next_metavar_orig = *next_metavar;
+                    let op = seq.kleene.op;
+                    let idx_first = locs.len();
+                    let idx_seq = idx_first - 1;
+                    inner(sess, &seq.tts, locs, next_metavar, seq_depth + 1);
+
+                    if let Some(separator) = &seq.separator {
+                        locs.push(MatcherLoc::SequenceSep { separator: separator.clone() });
+                        locs.push(MatcherLoc::SequenceKleeneOpAfterSep { idx_first });
+                    } else {
+                        locs.push(MatcherLoc::SequenceKleeneOpNoSep { op, idx_first });
+                    }
+
+                    // Overwrite the dummy value pushed above with the proper value.
+                    locs[idx_seq] = MatcherLoc::Sequence {
+                        op,
+                        num_metavar_decls: seq.num_captures,
+                        idx_first_after: locs.len(),
+                        next_metavar: next_metavar_orig,
+                        seq_depth,
+                    };
+                }
+                &TokenTree::MetaVarDecl(span, bind, kind) => {
+                    locs.push(MatcherLoc::MetaVarDecl {
+                        span,
+                        bind,
+                        kind,
+                        next_metavar: *next_metavar,
+                        seq_depth,
+                    });
+                    *next_metavar += 1;
+                }
+                TokenTree::MetaVar(..) | TokenTree::MetaVarExpr(..) => unreachable!(),
+            }
+        }
+    }
+
+    let mut locs = vec![];
+    let mut next_metavar = 0;
+    inner(sess, matcher, &mut locs, &mut next_metavar, /* seq_depth */ 0);
+
+    // A final entry is needed for eof.
+    locs.push(MatcherLoc::Eof);
+
+    locs
+}
+
 /// A single matcher position, representing the state of matching.
 struct MatcherPos {
     /// The index into `TtParser::locs`, which represents the "dot".
@@ -301,9 +373,6 @@ fn token_name_eq(t1: &Token, t2: &Token) -> bool {
 pub struct TtParser {
     macro_name: Ident,
 
-    /// The matcher of the current rule.
-    locs: Vec<MatcherLoc>,
-
     /// The set of current mps to be processed. This should be empty by the end of a successful
     /// execution of `parse_tt_inner`.
     cur_mps: Vec<MatcherPos>,
@@ -324,92 +393,13 @@ impl TtParser {
     pub(super) fn new(macro_name: Ident) -> TtParser {
         TtParser {
             macro_name,
-            locs: vec![],
             cur_mps: vec![],
             next_mps: vec![],
             bb_mps: vec![],
             empty_matches: Lrc::new(smallvec![]),
         }
     }
 
-    /// Convert a `&[TokenTree]` to a `&[MatcherLoc]`. Note: this conversion happens every time the
-    /// macro is called, which may be many times if there are many call sites or if it is
-    /// recursive. This conversion is fairly cheap and the representation is sufficiently better
-    /// for matching than `&[TokenTree]` that it's a clear performance win even with the overhead.
-    /// But it might be possible to move the conversion outwards so it only occurs once per macro.
-    fn compute_locs(&mut self, sess: &ParseSess, matcher: &[TokenTree]) -> usize {
-        fn inner(
-            sess: &ParseSess,
-            tts: &[TokenTree],
-            locs: &mut Vec<MatcherLoc>,
-            next_metavar: &mut usize,
-            seq_depth: usize,
-        ) {
-            for tt in tts {
-                match tt {
-                    TokenTree::Token(token) => {
-                        locs.push(MatcherLoc::Token { token: token.clone() });
-                    }
-                    TokenTree::Delimited(_, delimited) => {
-                        locs.push(MatcherLoc::Delimited);
-                        inner(sess, &delimited.all_tts, locs, next_metavar, seq_depth);
-                    }
-                    TokenTree::Sequence(_, seq) => {
-                        // We can't determine `idx_first_after` and construct the final
-                        // `MatcherLoc::Sequence` until after `inner()` is called and the sequence
-                        // end pieces are processed. So we push a dummy value (`Eof` is cheapest to
-                        // construct) now, and overwrite it with the proper value below.
-                        let dummy = MatcherLoc::Eof;
-                        locs.push(dummy);
-
-                        let next_metavar_orig = *next_metavar;
-                        let op = seq.kleene.op;
-                        let idx_first = locs.len();
-                        let idx_seq = idx_first - 1;
-                        inner(sess, &seq.tts, locs, next_metavar, seq_depth + 1);
-
-                        if let Some(separator) = &seq.separator {
-                            locs.push(MatcherLoc::SequenceSep { separator: separator.clone() });
-                            locs.push(MatcherLoc::SequenceKleeneOpAfterSep { idx_first });
-                        } else {
-                            locs.push(MatcherLoc::SequenceKleeneOpNoSep { op, idx_first });
-                        }
-
-                        // Overwrite the dummy value pushed above with the proper value.
-                        locs[idx_seq] = MatcherLoc::Sequence {
-                            op,
-                            num_metavar_decls: seq.num_captures,
-                            idx_first_after: locs.len(),
-                            next_metavar: next_metavar_orig,
-                            seq_depth,
-                        };
-                    }
-                    &TokenTree::MetaVarDecl(span, bind, kind) => {
-                        locs.push(MatcherLoc::MetaVarDecl {
-                            span,
-                            bind,
-                            kind,
-                            next_metavar: *next_metavar,
-                            seq_depth,
-                        });
-                        *next_metavar += 1;
-                    }
-                    TokenTree::MetaVar(..) | TokenTree::MetaVarExpr(..) => unreachable!(),
-                }
-            }
-        }
-
-        self.locs.clear();
-        let mut next_metavar = 0;
-        inner(sess, matcher, &mut self.locs, &mut next_metavar, /* seq_depth */ 0);
-
-        // A final entry is needed for eof.
-        self.locs.push(MatcherLoc::Eof);
-
-        // This is the number of metavar decls.
-        next_metavar
-    }
-
     /// Process the matcher positions of `cur_mps` until it is empty. In the process, this will
     /// produce more mps in `next_mps` and `bb_mps`.
     ///
@@ -420,15 +410,15 @@ impl TtParser {
     fn parse_tt_inner(
         &mut self,
         sess: &ParseSess,
-        num_metavar_decls: usize,
+        matcher: &[MatcherLoc],
         token: &Token,
     ) -> Option<NamedParseResult> {
         // Matcher positions that would be valid if the macro invocation was over now. Only
         // modified if `token == Eof`.
         let mut eof_mps = EofMatcherPositions::None;
 
         while let Some(mut mp) = self.cur_mps.pop() {
-            match &self.locs[mp.idx] {
+            match &matcher[mp.idx] {
                 MatcherLoc::Token { token: t } => {
                     // If it's a doc comment, we just ignore it and move on to the next tt in the
                     // matcher. This is a bug, but #95267 showed that existing programs rely on
@@ -536,7 +526,7 @@ impl TtParser {
                 }
                 MatcherLoc::Eof => {
                     // We are past the matcher's end, and not in a sequence. Try to end things.
-                    debug_assert_eq!(mp.idx, self.locs.len() - 1);
+                    debug_assert_eq!(mp.idx, matcher.len() - 1);
                     if *token == token::Eof {
                         eof_mps = match eof_mps {
                             EofMatcherPositions::None => EofMatcherPositions::One(mp),
@@ -554,11 +544,10 @@ impl TtParser {
         if *token == token::Eof {
             Some(match eof_mps {
                 EofMatcherPositions::One(mut eof_mp) => {
-                    assert_eq!(eof_mp.matches.len(), num_metavar_decls);
                     // Need to take ownership of the matches from within the `Lrc`.
                     Lrc::make_mut(&mut eof_mp.matches);
                     let matches = Lrc::try_unwrap(eof_mp.matches).unwrap().into_iter();
-                    self.nameize(sess, matches)
+                    self.nameize(sess, matcher, matches)
                 }
                 EofMatcherPositions::Multiple => {
                     Error(token.span, "ambiguity: multiple successful parses".to_string())
@@ -580,10 +569,8 @@ impl TtParser {
     pub(super) fn parse_tt(
         &mut self,
         parser: &mut Cow<'_, Parser<'_>>,
-        matcher: &[TokenTree],
+        matcher: &[MatcherLoc],
     ) -> NamedParseResult {
-        let num_metavar_decls = self.compute_locs(parser.sess, matcher);
-
         // A queue of possible matcher positions. We initialize it with the matcher position in
         // which the "dot" is before the first token of the first token tree in `matcher`.
         // `parse_tt_inner` then processes all of these possible matcher positions and produces
@@ -598,7 +585,7 @@ impl TtParser {
 
             // Process `cur_mps` until either we have finished the input or we need to get some
             // parsing from the black-box parser done.
-            if let Some(res) = self.parse_tt_inner(&parser.sess, num_metavar_decls, &parser.token) {
+            if let Some(res) = self.parse_tt_inner(&parser.sess, matcher, &parser.token) {
                 return res;
             }
 
@@ -626,7 +613,7 @@ impl TtParser {
                 (0, 1) => {
                     // We need to call the black-box parser to get some nonterminal.
                     let mut mp = self.bb_mps.pop().unwrap();
-                    let loc = &self.locs[mp.idx];
+                    let loc = &matcher[mp.idx];
                     if let &MatcherLoc::MetaVarDecl {
                         span,
                         kind: Some(kind),
@@ -664,19 +651,23 @@ impl TtParser {
 
                 (_, _) => {
                     // Too many possibilities!
-                    return self.ambiguity_error(parser.token.span);
+                    return self.ambiguity_error(matcher, parser.token.span);
                 }
             }
 
             assert!(!self.cur_mps.is_empty());
         }
     }
 
-    fn ambiguity_error(&self, token_span: rustc_span::Span) -> NamedParseResult {
+    fn ambiguity_error(
+        &self,
+        matcher: &[MatcherLoc],
+        token_span: rustc_span::Span,
+    ) -> NamedParseResult {
         let nts = self
             .bb_mps
             .iter()
-            .map(|mp| match &self.locs[mp.idx] {
+            .map(|mp| match &matcher[mp.idx] {
                 MatcherLoc::MetaVarDecl { bind, kind: Some(kind), .. } => {
                     format!("{} ('{}')", kind, bind)
                 }
@@ -702,12 +693,13 @@ impl TtParser {
     fn nameize<I: Iterator<Item = NamedMatch>>(
         &self,
         sess: &ParseSess,
+        matcher: &[MatcherLoc],
         mut res: I,
     ) -> NamedParseResult {
         // Make that each metavar has _exactly one_ binding. If so, insert the binding into the
         // `NamedParseResult`. Otherwise, it's an error.
         let mut ret_val = FxHashMap::default();
-        for loc in self.locs.iter() {
+        for loc in matcher {
             if let &MatcherLoc::MetaVarDecl { span, bind, kind, .. } = loc {
                 if kind.is_some() {
                     match ret_val.entry(MacroRulesNormalizedIdent::new(bind)) {

compiler/rustc_expand/src/mbe/macro_rules.rs

@@ -4,7 +4,7 @@ use crate::expand::{ensure_complete_parse, parse_ast_fragment, AstFragment, AstF
 use crate::mbe;
 use crate::mbe::macro_check;
 use crate::mbe::macro_parser::{Error, ErrorReported, Failure, Success, TtParser};
-use crate::mbe::macro_parser::{MatchedSeq, MatchedTokenTree};
+use crate::mbe::macro_parser::{MatchedSeq, MatchedTokenTree, MatcherLoc};
 use crate::mbe::transcribe::transcribe;
 
 use rustc_ast as ast;
@@ -159,7 +159,7 @@ struct MacroRulesMacroExpander {
     name: Ident,
     span: Span,
     transparency: Transparency,
-    lhses: Vec<mbe::TokenTree>,
+    lhses: Vec<Vec<MatcherLoc>>,
     rhses: Vec<mbe::TokenTree>,
     valid: bool,
     is_local: bool,
@@ -210,7 +210,7 @@ fn generic_extension<'cx, 'tt>(
     name: Ident,
     transparency: Transparency,
     arg: TokenStream,
-    lhses: &'tt [mbe::TokenTree],
+    lhses: &'tt [Vec<MatcherLoc>],
     rhses: &'tt [mbe::TokenTree],
     is_local: bool,
 ) -> Box<dyn MacResult + 'cx> {
@@ -245,14 +245,6 @@ fn generic_extension<'cx, 'tt>(
     // this situation.)
     let parser = parser_from_cx(sess, arg.clone());
 
-    // A matcher is always delimited, but the delimiters are ignored.
-    let delimited_inner_tts = |tt: &'tt mbe::TokenTree| -> &'tt [mbe::TokenTree] {
-        match tt {
-            mbe::TokenTree::Delimited(_, delimited) => delimited.inner_tts(),
-            _ => cx.span_bug(sp, "malformed macro lhs"),
-        }
-    };
-
     // Try each arm's matchers.
     let mut tt_parser = TtParser::new(name);
     for (i, lhs) in lhses.iter().enumerate() {
@@ -262,13 +254,19 @@ fn generic_extension<'cx, 'tt>(
         // are not recorded. On the first `Success(..)`ful matcher, the spans are merged.
         let mut gated_spans_snapshot = mem::take(&mut *sess.gated_spans.spans.borrow_mut());
 
-        match tt_parser.parse_tt(&mut Cow::Borrowed(&parser), delimited_inner_tts(lhs)) {
+        match tt_parser.parse_tt(&mut Cow::Borrowed(&parser), lhs) {
             Success(named_matches) => {
                 // The matcher was `Success(..)`ful.
                 // Merge the gated spans from parsing the matcher with the pre-existing ones.
                 sess.gated_spans.merge(gated_spans_snapshot);
 
-                let rhs = delimited_inner_tts(&rhses[i]).to_vec().clone();
+                // Ignore the delimiters on the RHS.
+                let rhs = match &rhses[i] {
+                    mbe::TokenTree::Delimited(_, delimited) => {
+                        delimited.inner_tts().to_vec().clone()
+                    }
+                    _ => cx.span_bug(sp, "malformed macro rhs"),
+                };
                 let arm_span = rhses[i].span();
 
                 let rhs_spans = rhs.iter().map(|t| t.span()).collect::<Vec<_>>();
@@ -346,10 +344,8 @@ fn generic_extension<'cx, 'tt>(
     // Check whether there's a missing comma in this macro call, like `println!("{}" a);`
     if let Some((arg, comma_span)) = arg.add_comma() {
         for lhs in lhses {
-            if let Success(_) = tt_parser.parse_tt(
-                &mut Cow::Borrowed(&parser_from_cx(sess, arg.clone())),
-                delimited_inner_tts(lhs),
-            ) {
+            let parser = parser_from_cx(sess, arg.clone());
+            if let Success(_) = tt_parser.parse_tt(&mut Cow::Borrowed(&parser), lhs) {
                 if comma_span.is_dummy() {
                     err.note("you might be missing a comma");
                 } else {
@@ -440,6 +436,8 @@ pub fn compile_declarative_macro(
             }),
         ),
     ];
+    // Convert it into `MatcherLoc` form.
+    let argument_gram = mbe::macro_parser::compute_locs(&sess.parse_sess, &argument_gram);
 
     let parser = Parser::new(&sess.parse_sess, body, true, rustc_parse::MACRO_ARGUMENTS);
     let mut tt_parser = TtParser::new(def.ident);
@@ -536,6 +534,25 @@ pub fn compile_declarative_macro(
         None => {}
     }
 
+    // Convert the lhses into `MatcherLoc` form, which is better for doing the
+    // actual matching. Unless the matcher is invalid.
+    let lhses = if valid {
+        lhses
+            .iter()
+            .map(|lhs| {
+                // Ignore the delimiters around the matcher.
+                match lhs {
+                    mbe::TokenTree::Delimited(_, delimited) => {
+                        mbe::macro_parser::compute_locs(&sess.parse_sess, delimited.inner_tts())
+                    }
+                    _ => sess.parse_sess.span_diagnostic.span_bug(def.span, "malformed macro lhs"),
+                }
+            })
+            .collect()
+    } else {
+        vec![]
+    };
+
     mk_syn_ext(Box::new(MacroRulesMacroExpander {
         name: def.ident,
         span: def.span,