Merge #7994

7994: Speed up mbe matching in heavy recursive cases r=edwin0cheng a=edwin0cheng

In some cases (e.g.  #4186), mbe matching is very slow due to a lot of copy and allocation for bindings, this PR try to solve this problem by introduce a semi "link-list" approach for bindings building.

I used this [test case](https://github.com/weiznich/minimal_example_for_rust_81262) (for `features(32-column-tables)`) to run following command to benchmark:
```
time rust-analyzer analysis-stats  --load-output-dirs ./ 
```

Before this PR : 2 mins
After this PR: 3 seconds.

However, for 64-column-tables cases, we still need 4 mins to complete. 

I will try to investigate in the following weeks.

bors r+




Co-authored-by: Edwin Cheng <[email protected]>

Author: bors[bot]

crates/hir_expand/src/db.rs

@@ -340,6 +340,8 @@ fn parse_macro_with_arg(
         None => return ExpandResult { value: None, err: result.err },
     };
 
+    log::debug!("expanded = {}", tt.as_debug_string());
+
     let fragment_kind = to_fragment_kind(db, macro_call_id);
 
     let (parse, rev_token_map) = match mbe::token_tree_to_syntax_node(&tt, fragment_kind) {

crates/mbe/src/expander.rs

@@ -5,7 +5,7 @@
 mod matcher;
 mod transcriber;
 
-use smallvec::SmallVec;
+use rustc_hash::FxHashMap;
 use syntax::SmolStr;
 
 use crate::{ExpandError, ExpandResult};
@@ -96,7 +96,7 @@ pub(crate) fn expand_rules(
 /// many is not a plain `usize`, but an `&[usize]`.
 #[derive(Debug, Default, Clone, PartialEq, Eq)]
 struct Bindings {
-    inner: SmallVec<[(SmolStr, Binding); 4]>,
+    inner: FxHashMap<SmolStr, Binding>,
 }
 
 #[derive(Debug, Clone, PartialEq, Eq)]