2026-02-24 23:44:07 -05:00
|
|
|
|
use crate::utils::Config;
|
2025-06-24 20:27:06 +02:00
|
|
|
|
use crossbeam_channel::{Receiver, Sender, bounded};
|
|
|
|
|
|
use ignore::{WalkBuilder, WalkState, overrides::OverrideBuilder};
|
2026-02-24 23:44:07 -05:00
|
|
|
|
use std::thread::JoinHandle;
|
2025-06-23 20:27:16 +02:00
|
|
|
|
use std::{
|
|
|
|
|
|
mem,
|
|
|
|
|
|
path::{Path, PathBuf},
|
|
|
|
|
|
thread,
|
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
// ---------------------------------------------------------------------------
|
|
|
|
|
|
// Internal constants / helpers
|
|
|
|
|
|
// ---------------------------------------------------------------------------
|
2025-06-16 23:47:50 +02:00
|
|
|
|
|
2026-02-24 23:44:07 -05:00
|
|
|
|
type Paths = Vec<PathBuf>;
|
2025-06-16 23:47:50 +02:00
|
|
|
|
|
2026-02-24 23:44:07 -05:00
|
|
|
|
struct BatchSender {
|
|
|
|
|
|
tx: Sender<Paths>,
|
|
|
|
|
|
batch: Paths,
|
|
|
|
|
|
batch_size: usize,
|
2025-06-16 23:47:50 +02:00
|
|
|
|
}
|
2026-02-24 23:44:07 -05:00
|
|
|
|
impl BatchSender {
|
|
|
|
|
|
fn new(tx: Sender<Paths>, batch_size: usize) -> Self {
|
|
|
|
|
|
Self {
|
|
|
|
|
|
tx,
|
|
|
|
|
|
batch: Vec::with_capacity(batch_size),
|
|
|
|
|
|
batch_size,
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
fn push_path(&mut self, path: PathBuf) {
|
|
|
|
|
|
self.batch.push(path);
|
|
|
|
|
|
if self.batch.len() >= self.batch_size {
|
2025-06-16 23:47:50 +02:00
|
|
|
|
self.flush();
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2026-02-24 23:44:07 -05:00
|
|
|
|
|
2025-06-16 23:47:50 +02:00
|
|
|
|
fn flush(&mut self) {
|
|
|
|
|
|
if !self.batch.is_empty() {
|
2026-02-24 23:44:07 -05:00
|
|
|
|
tracing::debug!(n_paths = self.batch.len(), "flushing batch");
|
2025-06-23 20:27:16 +02:00
|
|
|
|
let _ = self.tx.send(mem::take(&mut self.batch));
|
2025-06-16 23:47:50 +02:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2026-02-24 23:44:07 -05:00
|
|
|
|
impl Drop for BatchSender {
|
2025-06-24 20:27:06 +02:00
|
|
|
|
fn drop(&mut self) {
|
|
|
|
|
|
self.flush();
|
|
|
|
|
|
}
|
2025-06-16 23:47:50 +02:00
|
|
|
|
}
|
|
|
|
|
|
|
2026-02-24 23:44:07 -05:00
|
|
|
|
fn build_overrides(root: &Path, cfg: &Config) -> ignore::overrides::Override {
|
2025-06-16 23:47:50 +02:00
|
|
|
|
let mut ob = OverrideBuilder::new(root);
|
2026-02-24 23:44:07 -05:00
|
|
|
|
|
2025-06-16 23:47:50 +02:00
|
|
|
|
for ext in &cfg.scanner.excluded_extensions {
|
2025-06-17 02:22:14 +02:00
|
|
|
|
if let Err(e) = ob.add(&format!("!*.{ext}")) {
|
2026-02-24 23:44:07 -05:00
|
|
|
|
tracing::warn!("invalid exclude‐extension pattern ‘{ext}’: {e}");
|
2025-06-17 02:22:14 +02:00
|
|
|
|
}
|
2025-06-16 23:47:50 +02:00
|
|
|
|
}
|
|
|
|
|
|
for dir in &cfg.scanner.excluded_directories {
|
2025-06-17 02:22:14 +02:00
|
|
|
|
if let Err(e) = ob.add(&format!("!**/{dir}/**")) {
|
2026-02-24 23:44:07 -05:00
|
|
|
|
tracing::warn!("invalid exclude‐dir pattern ‘{dir}’: {e}");
|
2025-06-17 02:22:14 +02:00
|
|
|
|
}
|
2025-06-16 23:47:50 +02:00
|
|
|
|
}
|
2026-02-25 04:02:11 -05:00
|
|
|
|
for file in &cfg.scanner.excluded_files {
|
|
|
|
|
|
if let Err(e) = ob.add(&format!("!{file}")) {
|
|
|
|
|
|
tracing::warn!("invalid exclude‐file pattern ‘{file}’: {e}");
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2026-02-24 23:44:07 -05:00
|
|
|
|
|
|
|
|
|
|
ob.build().unwrap_or_else(|e| {
|
|
|
|
|
|
tracing::error!("failed to build ignore overrides: {e}");
|
|
|
|
|
|
ignore::overrides::Override::empty()
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
// ---------------------------------------------------------------------------
|
|
|
|
|
|
/// Walk `root` and send *batches* of paths through the returned channel.
|
|
|
|
|
|
pub fn spawn_file_walker(root: &Path, cfg: &Config) -> (Receiver<Paths>, JoinHandle<()>) {
|
|
|
|
|
|
let _span = tracing::info_span!("spawn_file_walker", root = %root.display()).entered();
|
|
|
|
|
|
let overrides = build_overrides(root, cfg);
|
2025-06-23 20:27:16 +02:00
|
|
|
|
|
|
|
|
|
|
// ----- 2 channel & thread pool parameters -----------------------------
|
2025-06-24 20:27:06 +02:00
|
|
|
|
let workers = cfg.performance.worker_threads.unwrap_or(num_cpus::get());
|
2026-02-24 23:44:07 -05:00
|
|
|
|
let (tx, rx) = bounded::<Paths>(workers * cfg.performance.channel_multiplier);
|
2025-06-16 23:47:50 +02:00
|
|
|
|
|
2025-06-24 20:27:06 +02:00
|
|
|
|
let root = root.to_path_buf();
|
2025-06-23 20:27:16 +02:00
|
|
|
|
let scan_hidden = cfg.scanner.scan_hidden_files;
|
2025-06-24 20:27:06 +02:00
|
|
|
|
let follow = cfg.scanner.follow_symlinks;
|
|
|
|
|
|
let max_bytes = cfg.scanner.max_file_size_mb.unwrap_or(0) * 1_048_576;
|
|
|
|
|
|
let batch_size = cfg.performance.batch_size;
|
2026-02-25 04:02:11 -05:00
|
|
|
|
let max_depth = cfg.performance.max_depth;
|
|
|
|
|
|
let same_file_system = cfg.scanner.one_file_system;
|
|
|
|
|
|
let require_git = cfg.scanner.require_git_to_read_vcsignore;
|
2025-06-23 20:27:16 +02:00
|
|
|
|
|
|
|
|
|
|
// ----- 3 the background walker thread ---------------------------------
|
2026-02-24 23:44:07 -05:00
|
|
|
|
let handle = thread::spawn(move || {
|
|
|
|
|
|
tracing::info!(
|
|
|
|
|
|
root = ?root,
|
|
|
|
|
|
workers = workers,
|
|
|
|
|
|
scan_hidden = scan_hidden,
|
|
|
|
|
|
follow_links = follow,
|
|
|
|
|
|
max_bytes = max_bytes,
|
|
|
|
|
|
batch_size = batch_size,
|
|
|
|
|
|
"starting directory walk"
|
|
|
|
|
|
);
|
|
|
|
|
|
|
2026-02-25 04:02:11 -05:00
|
|
|
|
let mut builder = WalkBuilder::new(root);
|
|
|
|
|
|
builder
|
2025-06-24 20:27:06 +02:00
|
|
|
|
.hidden(!scan_hidden)
|
|
|
|
|
|
.follow_links(follow)
|
|
|
|
|
|
.threads(workers)
|
|
|
|
|
|
.overrides(overrides)
|
2026-02-25 04:02:11 -05:00
|
|
|
|
.same_file_system(same_file_system)
|
|
|
|
|
|
.require_git(require_git);
|
|
|
|
|
|
if let Some(depth) = max_depth {
|
|
|
|
|
|
builder.max_depth(Some(depth));
|
|
|
|
|
|
}
|
|
|
|
|
|
builder
|
2026-02-24 23:44:07 -05:00
|
|
|
|
.filter_entry(|e| {
|
|
|
|
|
|
e.file_type()
|
|
|
|
|
|
.map(|ft| ft.is_dir() || ft.is_file())
|
|
|
|
|
|
.unwrap_or(true)
|
|
|
|
|
|
})
|
2025-06-24 20:27:06 +02:00
|
|
|
|
.build_parallel()
|
|
|
|
|
|
.run(move || {
|
2026-02-24 23:44:07 -05:00
|
|
|
|
let mut bs = BatchSender::new(tx.clone(), batch_size);
|
2025-06-23 18:25:10 +02:00
|
|
|
|
|
2025-06-24 20:27:06 +02:00
|
|
|
|
Box::new(move |entry| {
|
2026-02-24 23:44:07 -05:00
|
|
|
|
if let Ok(e) = entry {
|
|
|
|
|
|
let is_file = e.file_type().is_some_and(|ft| ft.is_file());
|
|
|
|
|
|
let under_limit = max_bytes == 0
|
|
|
|
|
|
|| e.metadata().map(|m| m.len() <= max_bytes).unwrap_or(true);
|
|
|
|
|
|
|
|
|
|
|
|
if is_file && under_limit {
|
|
|
|
|
|
bs.push_path(e.into_path());
|
2025-06-24 20:27:06 +02:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
WalkState::Continue
|
|
|
|
|
|
})
|
|
|
|
|
|
});
|
2026-02-24 23:44:07 -05:00
|
|
|
|
tracing::info!("directory walk complete");
|
2025-06-16 23:47:50 +02:00
|
|
|
|
});
|
2025-06-23 20:27:16 +02:00
|
|
|
|
|
2026-02-24 23:44:07 -05:00
|
|
|
|
(rx, handle)
|
2025-06-16 23:47:50 +02:00
|
|
|
|
}
|
2025-06-24 23:38:32 +02:00
|
|
|
|
|
|
|
|
|
|
#[test]
|
|
|
|
|
|
fn walker_respects_excluded_extensions() {
|
|
|
|
|
|
let tmp = tempfile::tempdir().unwrap();
|
|
|
|
|
|
std::fs::write(tmp.path().join("keep.rs"), "fn main(){}").unwrap();
|
|
|
|
|
|
std::fs::write(tmp.path().join("skip.txt"), "ignored").unwrap();
|
|
|
|
|
|
|
|
|
|
|
|
let mut cfg = Config::default();
|
|
|
|
|
|
cfg.scanner.excluded_extensions = vec!["txt".into()];
|
|
|
|
|
|
cfg.performance.worker_threads = Some(1);
|
|
|
|
|
|
cfg.performance.channel_multiplier = 1;
|
|
|
|
|
|
cfg.performance.batch_size = 2;
|
|
|
|
|
|
|
2026-02-24 23:44:07 -05:00
|
|
|
|
let (rx, handle) = spawn_file_walker(tmp.path(), &cfg);
|
|
|
|
|
|
if let Err(err) = handle.join() {
|
|
|
|
|
|
tracing::error!("walker thread panicked: {:#?}", err);
|
|
|
|
|
|
}
|
2025-06-24 23:38:32 +02:00
|
|
|
|
|
|
|
|
|
|
let all: Vec<_> = rx.into_iter().flatten().collect();
|
|
|
|
|
|
|
|
|
|
|
|
assert!(all.iter().any(|p| p.ends_with("keep.rs")));
|
|
|
|
|
|
assert!(all.iter().all(|p| !p.ends_with("skip.txt")));
|
|
|
|
|
|
}
|