From e3ff2872701d9f1d6f4f5223641092ec19c10a7a Mon Sep 17 00:00:00 2001
From: Matthias Schiffer <mschiffer@universe-factory.net>
Date: Mon, 1 Nov 2021 00:16:41 +0100
Subject: runner: detect file conflicts in dependencies again

The new code will also detect conflicts between the rootfs and
individual dependencies.
---
 Cargo.lock                |  1 +
 crates/runner/Cargo.toml  |  3 ++-
 crates/runner/src/task.rs | 67 +++++++++++++++++++++++++++++++++++++++++++++--
 3 files changed, 68 insertions(+), 3 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index b252090..3a8373c 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -538,6 +538,7 @@ dependencies = [
  "tar",
  "tee_readwrite",
  "uds",
+ "walkdir",
 ]
 
 [[package]]
diff --git a/crates/runner/Cargo.toml b/crates/runner/Cargo.toml
index d8fe562..19ad124 100644
--- a/crates/runner/Cargo.toml
+++ b/crates/runner/Cargo.toml
@@ -10,6 +10,7 @@ edition = "2021"
 [dependencies]
 common = { path = "../common", package = "rebel-common" }
 
+bincode = "1.3.3"
 blake3 = { version = "1.0.0", features = ["traits-preview"] }
 capctl = "0.2.0"
 digest = "0.9.0"
@@ -21,4 +22,4 @@ serde_json = "1.0.62"
 tar = "0.4.32"
 tee_readwrite = "0.1.0"
 uds = "0.2.6"
-bincode = "1.3.3"
+walkdir = "2.3.2"
diff --git a/crates/runner/src/task.rs b/crates/runner/src/task.rs
index 941e5f1..174768c 100644
--- a/crates/runner/src/task.rs
+++ b/crates/runner/src/task.rs
@@ -1,8 +1,8 @@
 use std::{
-	collections::{BTreeMap, HashMap},
+	collections::{BTreeMap, HashMap, HashSet},
 	io::{self, BufWriter},
 	os::unix::prelude::CommandExt,
-	path::Path,
+	path::{Path, PathBuf},
 	process::{self, Command, Stdio},
 	time::Instant,
 };
@@ -18,6 +18,7 @@ use serde::Serialize;
 use tee_readwrite::{TeeReader, TeeWriter};
 
 use common::{error::*, string_hash::*, types::*};
+use walkdir::WalkDir;
 
 use super::{
 	jobserver::Jobserver,
@@ -126,6 +127,59 @@ fn init_task(input_hash: &InputHash, task: &Task) -> Result<fs::Mount> {
 	Ok(mount)
 }
 
+fn get_contents<P1: AsRef<Path>, P2: AsRef<Path>>(
+	path: P1,
+	prefix: P2,
+) -> Result<(HashSet<PathBuf>, HashSet<PathBuf>)> {
+	let mut dirs = HashSet::new();
+	let mut files = HashSet::new();
+
+	let root: PathBuf = Path::new("/").join(prefix.as_ref());
+
+	for result in WalkDir::new(path.as_ref()).min_depth(1).into_iter() {
+		let entry = result
+			.with_context(|| format!("Failed to list contents of directory {:?}", path.as_ref()))?;
+		let is_dir = entry.file_type().is_dir();
+		let entry_path = root.join(entry.into_path().strip_prefix(path.as_ref()).unwrap());
+		if is_dir {
+			dirs.insert(entry_path);
+		} else {
+			files.insert(entry_path);
+		}
+	}
+
+	dirs.insert(root);
+
+	Ok((dirs, files))
+}
+
+fn check_conflicts(
+	dirs1: &HashSet<PathBuf>,
+	files1: &HashSet<PathBuf>,
+	dirs2: &HashSet<PathBuf>,
+	files2: &HashSet<PathBuf>,
+) -> Result<()> {
+	let mut conflicts = Vec::new();
+
+	conflicts.extend(files1.intersection(files2));
+	conflicts.extend(dirs1.intersection(files2));
+	conflicts.extend(files1.intersection(dirs2));
+
+	if !conflicts.is_empty() {
+		let mut conflict_strings: Box<[_]> = conflicts
+			.into_iter()
+			.map(|path| path.to_string_lossy().to_string())
+			.collect();
+		conflict_strings.sort();
+		return Err(Error::new(format!(
+			"Found the following file conflicts in dependencies:\n{}",
+			conflict_strings.join("\n")
+		)));
+	}
+
+	Ok(())
+}
+
 fn init_task_rootfs(input_hash: &InputHash, depends: &DependMap) -> Result<Stack<fs::Mount>> {
 	let task_tmp_dir = paths::task_tmp_dir(input_hash);
 	let mount_target = paths::join(&[&task_tmp_dir, paths::TASK_TMP_ROOTFS_SUBDIR]);
@@ -143,6 +197,8 @@ fn init_task_rootfs(input_hash: &InputHash, depends: &DependMap) -> Result<Stack
 		.with_context(|| format!("Failed to bind mount rootfs to {:?}", mount_target))?,
 	);
 
+	let (mut dirs, mut files) = get_contents(&mount_target, "")?;
+
 	for (path, dep_hashes) in depends {
 		assert!(!dep_hashes.is_empty());
 
@@ -156,6 +212,13 @@ fn init_task_rootfs(input_hash: &InputHash, depends: &DependMap) -> Result<Stack
 		let dep_target = mount_target.clone() + path;
 		let dep_paths: Box<[_]> = dep_hashes.iter().map(paths::depend_dir).collect();
 
+		for dep in dep_paths.iter() {
+			let (dep_dirs, dep_files) = get_contents(dep, path)?;
+			check_conflicts(&dirs, &files, &dep_dirs, &dep_files)?;
+			dirs.extend(dep_dirs);
+			files.extend(dep_files);
+		}
+
 		let options = format!(
 			"xino=off,index=off,metacopy=off,lowerdir={lower}:{base}",
 			lower = dep_paths.join(":"),
-- 
cgit v1.2.3