fix(cli): don't walk the subdirectory twice when using the --ignore flag (#8040)

This commit reworks "collect_files" utility to accept "ignore" parameter
which allows to filter out files in a single iteration instead of walking
file tree second time to excude "ignored" files.

cli/fmt.rs

@@ -9,8 +9,6 @@
 
 use crate::colors;
 use crate::diff::diff;
-use crate::fs::canonicalize_path;
-use crate::fs::files_in_subtree;
 use crate::text_encoding;
 use deno_core::error::generic_error;
 use deno_core::error::AnyError;
@@ -25,6 +23,7 @@ use std::path::Path;
 use std::path::PathBuf;
 use std::sync::atomic::{AtomicUsize, Ordering};
 use std::sync::{Arc, Mutex};
+use walkdir::WalkDir;
 
 const BOM_CHAR: char = '\u{FEFF}';
 
@@ -40,14 +39,8 @@ pub async fn format(
   if args.len() == 1 && args[0].to_string_lossy() == "-" {
     return format_stdin(check);
   }
-  // collect all files provided.
+  // collect the files that are to be formatted
-  let mut target_files = collect_files(args)?;
+  let target_files = collect_files(args, exclude)?;
-  if !exclude.is_empty() {
-    // collect all files to be ignored
-    // and retain only files that should be formatted.
-    let ignore_files = collect_files(exclude)?;
-    target_files.retain(|f| !ignore_files.contains(&f));
-  }
   let config = get_config();
   if check {
     check_source_files(config, target_files).await
@@ -234,22 +227,41 @@ fn is_supported(path: &Path) -> bool {
 
 pub fn collect_files(
   files: Vec<PathBuf>,
+  mut ignore: Vec<PathBuf>,
 ) -> Result<Vec<PathBuf>, std::io::Error> {
   let mut target_files: Vec<PathBuf> = vec![];
 
+  // retain only the paths which exist and ignore the rest
+  ignore.retain(|i| i.exists());
+
   if files.is_empty() {
-    target_files.extend(files_in_subtree(
+    for entry in WalkDir::new(std::env::current_dir()?)
-      canonicalize_path(&std::env::current_dir()?)?,
+      .into_iter()
-      is_supported,
+      .filter_entry(|e| {
-    ));
+        !ignore.iter().any(|i| {
+          e.path()
+            .canonicalize()
+            .unwrap()
+            .starts_with(i.canonicalize().unwrap())
+        })
+      })
+    {
+      let entry_clone = entry?.clone();
+      if is_supported(entry_clone.path()) {
+        target_files.push(entry_clone.path().canonicalize()?)
+      }
+    }
   } else {
     for file in files {
-      if file.is_dir() {
+      for entry in WalkDir::new(file)
-        target_files
+        .into_iter()
-          .extend(files_in_subtree(canonicalize_path(&file)?, is_supported));
+        .filter_entry(|e| !ignore.iter().any(|i| e.path().starts_with(i)))
-      } else {
+      {
-        target_files.push(canonicalize_path(&file)?);
+        let entry_clone = entry?.clone();
-      };
+        if is_supported(entry_clone.path()) {
+          target_files.push(entry_clone.into_path().canonicalize()?)
+        }
+      }
     }
   }
 

cli/lint.rs

@@ -47,13 +47,7 @@ pub async fn lint_files(
   if args.len() == 1 && args[0].to_string_lossy() == "-" {
     return lint_stdin(json);
   }
-  let mut target_files = collect_files(args)?;
+  let target_files = collect_files(args, ignore)?;
-  if !ignore.is_empty() {
-    // collect all files to be ignored
-    // and retain only files that should be linted.
-    let ignore_files = collect_files(ignore)?;
-    target_files.retain(|f| !ignore_files.contains(&f));
-  }
   debug!("Found {} files", target_files.len());
   let target_files_len = target_files.len();