validate performance optimizations (#64)

Author: perryqh

Cargo.lock

@@ -1,6 +1,6 @@
 [package]
 name = "codeowners"
-version = "0.2.4"
+version = "0.2.5"
 edition = "2024"
 
 [profile.release]
@@ -14,7 +14,7 @@ clap = { version = "4.5.20", features = ["derive"] }
 clap_derive = "4.5.18"
 error-stack = "0.5.0"
 enum_dispatch = "0.3.13"
-fast-glob = "0.4.0"
+fast-glob = "1.0.0"
 glob = "0.3.2"
 ignore = "0.4.23"
 itertools = "0.14.0"

Cargo.toml

@@ -178,9 +178,7 @@ impl Project {
     }
 
     pub fn relative_path<'a>(&'a self, absolute_path: &'a Path) -> &'a Path {
-        absolute_path
-            .strip_prefix(&self.base_path)
-            .expect("Could not generate relative path")
+        absolute_path.strip_prefix(&self.base_path).unwrap_or(absolute_path)
     }
 
     pub fn get_team(&self, name: &str) -> Option<Team> {
@@ -197,3 +195,35 @@ impl Project {
         result
     }
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_vendored_gem_by_name_maps_all_gems() {
+        let vg1 = VendoredGem {
+            path: PathBuf::from("vendored/a"),
+            name: "a".to_string(),
+        };
+        let vg2 = VendoredGem {
+            path: PathBuf::from("vendored/b"),
+            name: "b".to_string(),
+        };
+        let project = Project {
+            base_path: PathBuf::from("."),
+            files: vec![],
+            packages: vec![],
+            vendored_gems: vec![vg1.clone(), vg2.clone()],
+            teams: vec![],
+            codeowners_file_path: PathBuf::from(".github/CODEOWNERS"),
+            directory_codeowner_files: vec![],
+            teams_by_name: HashMap::new(),
+        };
+
+        let map = project.vendored_gem_by_name();
+        assert_eq!(map.len(), 2);
+        assert_eq!(map.get("a").unwrap().name, vg1.name);
+        assert_eq!(map.get("b").unwrap().name, vg2.name);
+    }
+}

src/project.rs

@@ -1,11 +1,12 @@
 use std::{
     fs::File,
     path::{Path, PathBuf},
+    sync::{Arc, Mutex},
 };
 
 use error_stack::{Result, ResultExt};
 use fast_glob::glob_match;
-use ignore::WalkBuilder;
+use ignore::{WalkBuilder, WalkParallel, WalkState};
 use rayon::iter::{IntoParallelIterator, ParallelIterator};
 use tracing::{instrument, warn};
 
@@ -53,12 +54,29 @@ impl<'a> ProjectBuilder<'a> {
         let mut entry_types = Vec::with_capacity(INITIAL_VECTOR_CAPACITY);
         let mut builder = WalkBuilder::new(&self.base_path);
         builder.hidden(false);
-        let walkdir = builder.build();
+        let walk_parallel: WalkParallel = builder.build_parallel();
 
-        for entry in walkdir {
-            let entry = entry.change_context(Error::Io)?;
+        let collected = Arc::new(Mutex::new(Vec::with_capacity(INITIAL_VECTOR_CAPACITY)));
+        let collected_for_threads = Arc::clone(&collected);
+
+        walk_parallel.run(move || {
+            let collected = Arc::clone(&collected_for_threads);
+            Box::new(move |res| {
+                if let Ok(entry) = res {
+                    if let Ok(mut v) = collected.lock() {
+                        v.push(entry);
+                    }
+                }
+                WalkState::Continue
+            })
+        });
+
+        // Process sequentially with &mut self
+        let collected_entries = Arc::try_unwrap(collected).unwrap().into_inner().unwrap();
+        for entry in collected_entries {
             entry_types.push(self.build_entry_type(entry)?);
         }
+
         self.build_project_from_entry_types(entry_types)
     }
 
@@ -216,7 +234,10 @@ impl<'a> ProjectBuilder<'a> {
 }
 
 fn matches_globs(path: &Path, globs: &[String]) -> bool {
-    globs.iter().any(|glob| glob_match(glob, path.to_str().unwrap()))
+    match path.to_str() {
+        Some(s) => globs.iter().any(|glob| glob_match(glob, s)),
+        None => false,
+    }
 }
 
 fn ruby_package_owner(path: &Path) -> Result<Option<String>, Error> {
@@ -241,14 +262,11 @@ mod tests {
 
     #[test]
     fn test_matches_globs() {
-        // should fail because hidden directories are ignored by glob patterns unless explicitly included
         assert!(matches_globs(Path::new("script/.eslintrc.js"), &[OWNED_GLOB.to_string()]));
     }
 
     #[test]
     fn test_glob_match() {
-        // Exposes bug in glob-match https://github.com/devongovett/glob-match/issues/9
-        // should fail because hidden directories are ignored by glob patterns unless explicitly included
         assert!(glob_match(OWNED_GLOB, "script/.eslintrc.js"));
     }
 }