SCANPY-181 use filter to extract tars when possible

Author: Seppli11

src/pysonar_scanner/jre.py

@@ -135,8 +135,7 @@ def __extract_jre(self, file_path: pathlib.Path, unzip_dir: pathlib.Path):
             with zipfile.ZipFile(file_path, "r") as zip_ref:
                 zip_ref.extractall(unzip_dir)
         elif file_path.suffix in [".gz", ".tgz"]:
-            with tarfile.open(file_path, "r:gz") as tar_ref:
-                tar_ref.extractall(unzip_dir, filter="data")
+            utils.extract_tar(file_path, unzip_dir)
         else:
             raise UnsupportedArchiveFormat(
                 f"Received JRE is packaged as an unsupported archive format: {file_path.suffix}"

src/pysonar_scanner/utils.py

@@ -20,6 +20,8 @@
 import hashlib
 import pathlib
 import platform
+import sys
+import tarfile
 import typing
 from enum import Enum
 
@@ -89,3 +91,11 @@ def get_arch() -> Arch:
 
 def filter_none_values(dictionary: dict) -> dict:
     return {k: v for k, v in dictionary.items() if v is not None}
+
+
+def extract_tar(path: pathlib.Path, target_dir: pathlib.Path):
+    with tarfile.open(path, "r:gz") as tar_ref:
+        if sys.version_info >= (3, 12):
+            tar_ref.extractall(target_dir, filter="data")
+        else:
+            tar_ref.extractall(target_dir)

tests/unit/test_utils.py

@@ -23,7 +23,7 @@
 import unittest.mock
 import pyfakefs.fake_filesystem_unittest as pyfakefs
 
-from pysonar_scanner.utils import Arch, Os, get_arch, get_os, remove_trailing_slash, calculate_checksum
+from pysonar_scanner.utils import Arch, Os, get_arch, get_os, remove_trailing_slash, calculate_checksum, extract_tar
 
 
 class TestUtils(unittest.TestCase):
@@ -132,3 +132,31 @@ def test_calculate_checksum(self):
             calculate_checksum(BytesIO(b"test test")),
             "03ffdf45276dd38ffac79b0e9c6c14d89d9113ad783d5922580f4c66a3305591",
         )
+
+
+class TestExtractTar(unittest.TestCase):
+    def setUp(self):
+        self.test_path = pathlib.Path("/fake/path/archive.tar.gz")
+        self.test_target_dir = pathlib.Path("/fake/target/dir")
+
+    @unittest.mock.patch("tarfile.open")
+    @unittest.mock.patch("sys.version_info", (3, 12, 0))
+    def test_extract_tar_python_3_12_or_higher(self, mock_open):
+        mock_tar = unittest.mock.MagicMock()
+        mock_open.return_value.__enter__.return_value = mock_tar
+
+        extract_tar(self.test_path, self.test_target_dir)
+
+        mock_open.assert_called_once_with(self.test_path, "r:gz")
+        mock_tar.extractall.assert_called_once_with(self.test_target_dir, filter="data")
+
+    @unittest.mock.patch("tarfile.open")
+    @unittest.mock.patch("sys.version_info", (3, 11, 0))
+    def test_extract_tar_python_older_than_3_12(self, mock_open):
+        mock_tar = unittest.mock.MagicMock()
+        mock_open.return_value.__enter__.return_value = mock_tar
+
+        extract_tar(self.test_path, self.test_target_dir)
+
+        mock_open.assert_called_once_with(self.test_path, "r:gz")
+        mock_tar.extractall.assert_called_once_with(self.test_target_dir)