file_finder: Respect .gitignore and file_scan_inclusions with ** in glob (#40654)

Closes #39037 

Previously, the code split the `**/.env` glob in `file_scan_inclusions`
into two sources for the `PathMatcher`: `["**", "**/.env"]`. This
approach works for directories, but including `**` will match all
directories and their files. To address this, I now select the
appropriate `PathMatcher` using only `**/.env` when specifically
targeting a file to determine whether to include it in the file finder.

Release Notes:

- Fixed: respect `.gitignore` and `file_scan_inclusions` settings with
`**` in glob for file finder

---------

Signed-off-by: Benjamin <5719034+bnjjj@users.noreply.github.com>
Co-authored-by: Julia Ryan <juliaryan3.14@gmail.com>

crates/worktree/src/worktree.rs

@@ -505,7 +505,7 @@ impl Worktree {
                 project_id,
                 replica_id,
                 snapshot,
-                file_scan_inclusions: settings.file_scan_inclusions.clone(),
+                file_scan_inclusions: settings.parent_dir_scan_inclusions.clone(),
                 background_snapshot: background_snapshot.clone(),
                 updates_tx: Some(background_updates_tx),
                 update_observer: None,
@@ -520,8 +520,10 @@ impl Worktree {
                 while let Some(update) = background_updates_rx.next().await {
                     {
                         let mut lock = background_snapshot.lock();
-                        lock.0
-                            .apply_remote_update(update.clone(), &settings.file_scan_inclusions);
+                        lock.0.apply_remote_update(
+                            update.clone(),
+                            &settings.parent_dir_scan_inclusions,
+                        );
                         lock.1.push(update);
                     }
                     snapshot_updated_tx.send(()).await.ok();
@@ -4290,7 +4292,8 @@ impl BackgroundScanner {
 
             if child_entry.is_dir() {
                 child_entry.is_ignored = ignore_stack.is_abs_path_ignored(&child_abs_path, true);
-                child_entry.is_always_included = self.settings.is_path_always_included(&child_path);
+                child_entry.is_always_included =
+                    self.settings.is_path_always_included(&child_path, true);
 
                 // Avoid recursing until crash in the case of a recursive symlink
                 if job.ancestor_inodes.contains(&child_entry.inode) {
@@ -4315,7 +4318,8 @@ impl BackgroundScanner {
                 }
             } else {
                 child_entry.is_ignored = ignore_stack.is_abs_path_ignored(&child_abs_path, false);
-                child_entry.is_always_included = self.settings.is_path_always_included(&child_path);
+                child_entry.is_always_included =
+                    self.settings.is_path_always_included(&child_path, false);
             }
 
             {
@@ -4450,7 +4454,8 @@ impl BackgroundScanner {
                     fs_entry.is_ignored = ignore_stack.is_abs_path_ignored(&abs_path, is_dir);
                     fs_entry.is_external = is_external;
                     fs_entry.is_private = self.is_path_private(path);
-                    fs_entry.is_always_included = self.settings.is_path_always_included(path);
+                    fs_entry.is_always_included =
+                        self.settings.is_path_always_included(path, is_dir);
 
                     let parent_is_hidden = path
                         .parent()

crates/worktree/src/worktree_settings.rs

@@ -13,8 +13,11 @@ pub struct WorktreeSettings {
     pub project_name: Option<String>,
     /// Whether to prevent this project from being shared in public channels.
     pub prevent_sharing_in_public_channels: bool,
-    pub file_scan_inclusions: PathMatcher,
     pub file_scan_exclusions: PathMatcher,
+    pub file_scan_inclusions: PathMatcher,
+    /// This field contains all ancestors of the `file_scan_inclusions`. It's used to
+    /// determine whether to terminate worktree scanning for a given dir.
+    pub parent_dir_scan_inclusions: PathMatcher,
     pub private_files: PathMatcher,
 }
 
@@ -29,9 +32,12 @@ impl WorktreeSettings {
             .any(|ancestor| self.file_scan_exclusions.is_match(ancestor.as_std_path()))
     }
 
-    pub fn is_path_always_included(&self, path: &RelPath) -> bool {
-        path.ancestors()
-            .any(|ancestor| self.file_scan_inclusions.is_match(ancestor.as_std_path()))
+    pub fn is_path_always_included(&self, path: &RelPath, is_dir: bool) -> bool {
+        if is_dir {
+            self.parent_dir_scan_inclusions.is_match(path.as_std_path())
+        } else {
+            self.file_scan_inclusions.is_match(path.as_std_path())
+        }
     }
 }
 
@@ -46,6 +52,7 @@ impl Settings for WorktreeSettings {
             .flat_map(|glob| {
                 Path::new(glob)
                     .ancestors()
+                    .skip(1)
                     .map(|a| a.to_string_lossy().into())
             })
             .filter(|p: &String| !p.is_empty())
@@ -57,11 +64,13 @@ impl Settings for WorktreeSettings {
             file_scan_exclusions: path_matchers(file_scan_exclusions, "file_scan_exclusions")
                 .log_err()
                 .unwrap_or_default(),
-            file_scan_inclusions: path_matchers(
+            parent_dir_scan_inclusions: path_matchers(
                 parsed_file_scan_inclusions,
                 "file_scan_inclusions",
             )
             .unwrap(),
+            file_scan_inclusions: path_matchers(file_scan_inclusions, "file_scan_inclusions")
+                .unwrap(),
             private_files: path_matchers(private_files, "private_files")
                 .log_err()
                 .unwrap_or_default(),

crates/worktree/src/worktree_tests.rs

@@ -760,6 +760,7 @@ async fn test_file_scan_inclusions(cx: &mut TestAppContext) {
             "prettier": {
                 "package.json": "{}",
             },
+            "package.json": "//package.json"
         },
         "src": {
             ".DS_Store": "",