don't poison mutex around chalk

We use panics for cancellation, so we could trigger panic while
holding the solver. std::sync::Mutex will be poisoned as a result,
which and all further attempts to use solver (from other threads) will
panic as well.

This commit switches to parking_lot::Mutex which just unlocks on panic.

crates/ra_hir/src/db.rs

@@ -1,5 +1,6 @@
-use std::sync::{Arc, Mutex};
+use std::sync::Arc;
 
+use parking_lot::Mutex;
 use ra_syntax::{SyntaxNode, TreeArc, SmolStr, ast};
 use ra_db::{SourceDatabase, salsa};
 

crates/ra_hir/src/ty/traits.rs

@@ -1,6 +1,7 @@
 //! Trait solving using Chalk.
-use std::sync::{Arc, Mutex};
+use std::sync::Arc;
 
+use parking_lot::Mutex;
 use rustc_hash::FxHashSet;
 use log::debug;
 use chalk_ir::cast::Cast;
@@ -61,7 +62,7 @@ fn solve(
     let context = ChalkContext { db, krate };
     let solver = db.solver(krate);
     debug!("solve goal: {:?}", goal);
-    let solution = solver.lock().unwrap().solve_with_fuel(&context, goal, Some(1000));
+    let solution = solver.lock().solve_with_fuel(&context, goal, Some(1000));
     debug!("solve({:?}) => {:?}", goal, solution);
     solution
 }