Switch from Arc/RwLock to Rc/RefCell for CodeContextMenu (#22035)

`CodeContextMenu` is always accessed on one thread, so only `Rc`s and
`Rc<RefCell<_>>` are needed. There should be tiny performance benefits
from this. The main benefit of this is that when seeing code accessing a
`RwLock` it would be reasonable to wonder whether it will block. The
only potential downside is the potential for panics due to overlapping
borrows of the RefCells. I think this is an acceptable risk because most
errors of this nature will be local or will be caught by clippy via the
check for holding a RefCell reference over an `await`.

Release Notes:

- N/A

crates/project/src/project.rs

@@ -57,7 +57,7 @@ use lsp::{
 };
 use lsp_command::*;
 use node_runtime::NodeRuntime;
-use parking_lot::{Mutex, RwLock};
+use parking_lot::Mutex;
 pub use prettier_store::PrettierStore;
 use project_settings::{ProjectSettings, SettingsObserver, SettingsObserverEvent};
 use remote::{SshConnectionOptions, SshRemoteClient};
@@ -73,8 +73,10 @@ use snippet::Snippet;
 use snippet_provider::SnippetProvider;
 use std::{
     borrow::Cow,
+    cell::RefCell,
     ops::Range,
     path::{Component, Path, PathBuf},
+    rc::Rc,
     str,
     sync::Arc,
     time::Duration,
@@ -2872,7 +2874,7 @@ impl Project {
         &self,
         buffer: Model<Buffer>,
         completion_indices: Vec<usize>,
-        completions: Arc<RwLock<Box<[Completion]>>>,
+        completions: Rc<RefCell<Box<[Completion]>>>,
         cx: &mut ModelContext<Self>,
     ) -> Task<Result<bool>> {
         self.lsp_store.update(cx, |lsp_store, cx| {