Add skip_path feature

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "findlargedir"
-version = "0.1.3"
+version = "0.1.4"
 authors = ["Dinko Korunic <[email protected]>"]
 categories = ["command-line-utilities"]
 description = "find all blackhole directories with a huge amount of filesystem entries in a flat structure"

README.md

@@ -46,6 +46,9 @@ OPTIONS:
     -o, --one-filesystem <ONE_FILESYSTEM>
             [default: true] [possible values: true, false]
 
+    -s, --skip-path <SKIP_PATH>
+
+
     -t, --calibration-path <CALIBRATION_PATH>
 
 

src/args.rs

@@ -1,27 +1,39 @@
 use clap::Parser;
+use clap::ValueHint;
 use std::path::PathBuf;
 
-#[derive(Parser, Debug)]
+#[derive(Parser, Default, Debug)]
 #[clap(author, version, about, long_about = None)]
 pub struct Args {
+    // TODO: Perform accurate directory entry counting
     #[clap(short = 'a', long, action = clap::ArgAction::Set, default_value_t = false)]
     pub accurate: bool,
 
+    // Do not cross mount points
     #[clap(short = 'o', long, action = clap::ArgAction::Set, default_value_t = true)]
     pub one_filesystem: bool,
 
+    // Calibration entry count
     #[clap(short = 'c', long, value_parser, default_value_t = crate::calibrate::DEFAULT_TEST_COUNT)]
     pub calibration_count: u64,
 
-    #[clap(short = 't', long, value_parser)]
+    // Calibration path
+    #[clap(short = 't', long, value_parser, value_hint = ValueHint::AnyPath)]
     pub calibration_path: Option<PathBuf>,
 
+    // Alert threshold count: just print the estimate
     #[clap(short = 'A', long, value_parser, default_value_t = crate::walk::ALERT_COUNT)]
     pub alert_threshold: u64,
 
+    // Blacklist threshold count: print the estimate and stop further deeper scan
     #[clap(short = 'B', long, value_parser, default_value_t = crate::walk::BLACKLIST_COUNT)]
     pub blacklist_threshold: u64,
 
-    #[clap(required = true, value_parser)]
+    // Directories to never scan
+    #[clap(short = 's', long, value_parser, value_hint = ValueHint::AnyPath)]
+    pub skip_path: Vec<PathBuf>,
+
+    // Paths to perform deep scan on
+    #[clap(required = true, value_parser, value_hint = ValueHint::AnyPath)]
     pub path: Vec<PathBuf>,
 }

src/walk.rs

@@ -4,6 +4,7 @@ use fs_err as fs;
 use human_bytes::human_bytes;
 use human_format::Formatter;
 use jwalk::{DirEntry, Parallelism, WalkDir};
+use std::collections::HashSet;
 use std::fs::Metadata;
 use std::os::unix::fs::MetadataExt;
 use std::path::Path;
@@ -36,6 +37,7 @@ pub fn parallel_search(
         args.alert_threshold,
         args.blacklist_threshold,
     );
+    let skip_path = args.skip_path.iter().cloned().collect::<HashSet<_>>();
 
     for _ in WalkDir::new(path)
         .skip_hidden(false)
@@ -53,6 +55,7 @@ pub fn parallel_search(
                     &path_metadata,
                     size_inode_ratio,
                     dir_entry_result,
+                    &skip_path,
                     one_filesystem,
                     alert_threshold,
                     blacklist_threshold,
@@ -70,13 +73,25 @@ fn process_dir_entry<E>(
     path_metadata: &Metadata,
     size_inode_ratio: u64,
     dir_entry_result: &mut Result<DirEntry<((), ())>, E>,
+    skip_path: &HashSet<PathBuf>,
     one_filesystem: bool,
     alert_threshold: u64,
     blacklist_threshold: u64,
 ) {
     if let Ok(dir_entry) = dir_entry_result {
         if dir_entry.file_type.is_dir() {
             if let Some(full_path) = dir_entry.read_children_path.as_ref() {
+                // Ignore skip paths, typically being virtual filesystems (/proc, /dev, /sys, /run)
+                if !skip_path.is_empty() && skip_path.contains(&full_path.to_path_buf()) {
+                    println!(
+                        "Skipping further scan at {} as requested",
+                        full_path.display()
+                    );
+
+                    dir_entry.read_children_path = None;
+                    return;
+                }
+
                 // Retrieve Unix metadata for a given directory
                 if let Ok(dir_entry_metadata) = fs::metadata(full_path) {
                     // If `one_filesystem` flag has been set and if directory is not residing