fix: tar hardlink is not preserved after decompressing or compressing (#879)

* fix tar hard link compress and decompress

Signed-off-by: tommady <[email protected]>

* let tar can support windows

Signed-off-by: tommady <[email protected]>

* let tar can support windows

Signed-off-by: tommady <[email protected]>

* let tar not supporting windows

Signed-off-by: tommady <[email protected]>

* fix tar linter

Signed-off-by: tommady <[email protected]>

* add integration test

Signed-off-by: tommady <[email protected]>

* update CHANGELOG

Signed-off-by: tommady <[email protected]>

---------

Signed-off-by: tommady <[email protected]>

Author: tommady

CHANGELOG.md

@@ -46,6 +46,7 @@ Categories Used:
 - Handle broken symlinks in zip archives and normalize path separators [\#841](https://github.com/ouch-org/ouch/pull/841) ([zzzsyyy](https://github.com/zzzsyyy))
 - Fix folder softlink is not preserved after packing [\#850](https://github.com/ouch-org/ouch/pull/850) ([tommady](https://github.com/tommady))
 - Handle read-only directories in tar extraction [\#873](https://github.com/ouch-org/ouch/pull/873) ([vrmiguel](https://github.com/vrmiguel))
+- Fix tar hardlink is not preserved after decompressing or compressing [\#879](https://github.com/ouch-org/ouch/pull/879) ([tommady](https://github.com/tommady))
 - Fix enable gitignore flag should work without git [\#881](https://github.com/ouch-org/ouch/pull/881) ([tommady](https://github.com/tommady))
 
 ### Tweaks

Cargo.lock

@@ -1,6 +1,9 @@
 //! Contains Tar-specific building and unpacking functions
 
+#[cfg(unix)]
+use std::os::unix::fs::MetadataExt;
 use std::{
+    collections::HashMap,
     env,
     io::prelude::*,
     ops::Not,
@@ -42,6 +45,17 @@ pub fn unpack_archive(reader: Box<dyn Read>, output_folder: &Path, quiet: bool)
 
                 create_symlink(&target, &full_path)?;
             }
+            tar::EntryType::Link => {
+                let link_path = file.path()?;
+                let target = file
+                    .link_name()?
+                    .ok_or_else(|| std::io::Error::new(std::io::ErrorKind::InvalidData, "Missing hardlink target"))?;
+
+                let full_link_path = output_folder.join(&link_path);
+                let full_target_path = output_folder.join(&target);
+
+                std::fs::hard_link(&full_target_path, &full_link_path)?;
+            }
             tar::EntryType::Regular => {
                 file.unpack_in(output_folder)?;
             }
@@ -135,6 +149,7 @@ where
 {
     let mut builder = tar::Builder::new(writer);
     let output_handle = Handle::from_path(output_path);
+    let mut seen_inode: HashMap<(u64, u64), PathBuf> = HashMap::new();
 
     for filename in input_filenames {
         let previous_location = utils::cd_into_same_dir_as(filename)?;
@@ -164,7 +179,9 @@ where
                 info!("Compressing '{}'", EscapedPathDisplay::new(path));
             }
 
-            if !follow_symlinks && path.symlink_metadata()?.is_symlink() {
+            let link_meta = path.symlink_metadata()?;
+
+            if !follow_symlinks && link_meta.is_symlink() {
                 let target_path = path.read_link()?;
 
                 let mut header = tar::Header::new_gnu();
@@ -176,25 +193,59 @@ where
                         .detail("Unexpected error while trying to read link")
                         .detail(format!("Error: {err}."))
                 })?;
-            } else if path.is_dir() {
-                builder.append_dir(path, path)?;
-            } else {
-                let mut file = match fs::File::open(path) {
-                    Ok(f) => f,
-                    Err(e) => {
-                        if e.kind() == std::io::ErrorKind::NotFound && path.is_symlink() {
-                            // This path is for a broken symlink, ignore it
-                            continue;
-                        }
-                        return Err(e.into());
+                continue;
+            }
+
+            // TODO: for supporting windows hard link easier
+            // we should wait for this issue
+            // https://github.com/rust-lang/rust/issues/63010
+            #[cfg(unix)]
+            if link_meta.nlink() > 1 && link_meta.is_file() {
+                let key = (link_meta.dev(), link_meta.ino());
+
+                match seen_inode.get(&key) {
+                    Some(target_path) => {
+                        let mut header = tar::Header::new_gnu();
+                        header.set_entry_type(tar::EntryType::Link);
+                        header.set_size(0);
+
+                        builder.append_link(&mut header, path, target_path).map_err(|err| {
+                            FinalError::with_title("Could not create archive").detail(format!(
+                                "Error appending hard link '{}': {}",
+                                path.display(),
+                                err
+                            ))
+                        })?;
                     }
-                };
-                builder.append_file(path, file.file_mut()).map_err(|err| {
-                    FinalError::with_title("Could not create archive")
-                        .detail("Unexpected error while trying to read file")
-                        .detail(format!("Error: {err}."))
-                })?;
+                    None => {
+                        seen_inode.insert(key, path.to_path_buf());
+                        let mut file = fs::File::open(path)?;
+                        builder.append_file(path, file.file_mut())?
+                    }
+                }
+                continue;
+            }
+
+            if path.is_dir() {
+                builder.append_dir(path, path)?;
+                continue;
             }
+
+            let mut file = match fs::File::open(path) {
+                Ok(f) => f,
+                Err(e) => {
+                    if e.kind() == std::io::ErrorKind::NotFound && path.is_symlink() {
+                        // This path is for a broken symlink, ignore it
+                        continue;
+                    }
+                    return Err(e.into());
+                }
+            };
+            builder.append_file(path, file.file_mut()).map_err(|err| {
+                FinalError::with_title("Could not create archive")
+                    .detail("Unexpected error while trying to read file")
+                    .detail(format!("Error: {err}."))
+            })?;
         }
         env::set_current_dir(previous_location)?;
     }

src/archive/tar.rs

@@ -1148,3 +1148,64 @@ fn sevenz_list_should_not_failed() {
 
     assert!(res.get_output().stdout.find(b"README.md").is_some());
 }
+
+// TODO: for supporting windows hard link easier
+// we should wait for this issue
+// https://github.com/rust-lang/rust/issues/63010
+#[cfg(unix)]
+#[test]
+fn tar_hardlink_pack_and_unpack() {
+    use std::{fs::hard_link, os::unix::fs::MetadataExt};
+
+    let temp_dir = tempdir().unwrap();
+    let root_path = temp_dir.path();
+    let source_path = root_path.join("hardlink");
+    fs::create_dir_all(&source_path).unwrap();
+    let out_path = root_path.join("out");
+    fs::create_dir_all(&out_path).unwrap();
+
+    let source = fs::File::create(source_path.join("source")).unwrap();
+    let link1 = source_path.join("link1");
+    let link2 = source_path.join("link2");
+    hard_link(source.path(), link1.as_path()).unwrap();
+    hard_link(source.path(), link2.as_path()).unwrap();
+
+    let archive = root_path.join("archive.tar.gz");
+    crate::utils::cargo_bin()
+        .arg("compress")
+        .arg(&source_path)
+        .arg(&archive)
+        .assert()
+        .success();
+
+    crate::utils::cargo_bin()
+        .arg("decompress")
+        .arg(archive)
+        .arg("-d")
+        .arg(&out_path)
+        .assert()
+        .success();
+
+    let out_source_meta = fs::File::open(out_path.join("hardlink").join("source"))
+        .unwrap()
+        .metadata()
+        .unwrap();
+    let out_link1_meta = fs::File::open(out_path.join("hardlink").join("link1"))
+        .unwrap()
+        .metadata()
+        .unwrap();
+    let out_link2_meta = fs::File::open(out_path.join("hardlink").join("link2"))
+        .unwrap()
+        .metadata()
+        .unwrap();
+
+    assert!(out_source_meta.nlink() > 1);
+    assert!(out_link1_meta.nlink() > 1);
+    assert!(out_link2_meta.nlink() > 1);
+
+    assert_eq!(out_source_meta.dev(), out_link1_meta.dev());
+    assert_eq!(out_link1_meta.dev(), out_link2_meta.dev());
+
+    assert_eq!(out_source_meta.ino(), out_link1_meta.ino());
+    assert_eq!(out_link1_meta.ino(), out_link2_meta.ino());
+}