Make utils.extract_tar work

The same way as it was working with extractcode.

Signed-off-by: Philippe Ombredanne <pombredanne@nexb.com>

CHANGELOG.rst

@@ -6,10 +6,10 @@ v31.0.0
 
 This is a major release with bug fixes and API changes.
 
-- Remove dependency on extractcode. "extract_tar" has been removed too. Use
-  instead extract_tar_with_symlinks.
-- Adopted the latest skeleton. 
-- Add new os tests
+- Remove dependency on extractcode as images only need basic tar to extract.
+  - "utils.extract_tar" function now accepts a skip_symlink argument
+- Adopt the latest skeleton. 
+- Add new os-release tests
 
 
 v30.0.0 

src/container_inspector/image.py

@@ -450,14 +450,15 @@ def get_installed_packages(self, packages_getter):
                 yield purl, package, layer
 
     @staticmethod
-    def extract(archive_location, extracted_location):
+    def extract(archive_location, extracted_location, skip_symlinks=False):
         """
         Extract the image archive tarball at ``archive_location`` to
-        ``extracted_location``.
+        ``extracted_location``. Skip symlinks and links if ``skip_symlinks`` is True.
         """
-        utils.extract_tar_keeping_symlinks(
+        utils.extract_tar(
             location=archive_location,
             target_dir=extracted_location,
+            skip_symlinks=skip_symlinks,
         )
 
     @staticmethod
@@ -1070,15 +1071,16 @@ def __attrs_post_init__(self, *args, **kwargs):
         if not self.size:
             self.size = os.path.getsize(self.archive_location)
 
-    def extract(self, extracted_location):
+    def extract(self, extracted_location, skip_symlinks=True):
         """
         Extract this layer archive in the `extracted_location` directory and set
         this Layer ``extracted_location`` attribute to ``extracted_location``.
         """
         self.extracted_location = extracted_location
-        utils.extract_tar_keeping_symlinks(
+        utils.extract_tar(
             location=self.archive_location,
             target_dir=extracted_location,
+            skip_symlinks=skip_symlinks,
         )
 
     def get_resources(self, with_dir=False, walker=os.walk):

src/container_inspector/utils.py

@@ -80,40 +80,57 @@ def get_labels(config, container_config):
     return dict(sorted(labels.items()))
 
 
-def extract_tar(location, target_dir):
+def extract_tar(location, target_dir, skip_symlinks=True):
     """
     Extract a tar archive at `location` in the `target_dir` directory.
-    Ignore special device files and symlinks and hardlinks.
-    Do not preserve the permissions and owners.
-    Raise an Exception on error.
-    """
-    raise NotImplementedError('This function has been removed, use extract_tar_keeping_symlinks instead')
-
-
-def extract_tar_keeping_symlinks(location, target_dir):
-    """
-    Extract a tar archive at `location` in the `target_dir` directory.
-    Ignore special device files. Keep symlinks and hardlinks
+    Ignore special device files. Skip symlinks and hardlinks if skip_symlinks is True.
     Do not preserve the permissions and owners.
     Raise exceptions on possible problematic relative paths.
+    Issue a warning if skip_symlinks is True and links target are missing.
     """
     import tarfile
-    if TRACE: logger.debug(f'extract_tar_keeping_symlinks: {location} to {target_dir}')
+    tarfile.TarInfo
+    if TRACE: logger.debug(f'_extract_tar: {location} to {target_dir} skip_symlinks: {skip_symlinks}')
 
     fileutils.create_dir(target_dir)
 
     with tarfile.open(location) as tarball:
         # never extract character device, block and fifo files:
         # we extract dirs, files and links only
-        for tinfo in tarball:
-            if tinfo.isdev():
+        to_extract = []
+        for tarinfo in tarball:
+            if TRACE: logger.debug(f'_extract_tar: {tarinfo}')
+
+            if tarinfo.isdev() or tarinfo.ischr() or tarinfo.isblk() or tarinfo.isfifo() or tarinfo.sparse:
+                if TRACE:
+                    logger.debug(f'_extract_tar: skipping unsupported {tarinfo} file type: block, chr, dev or sparse file')
                 continue
-            if TRACE: logger.debug(f'extract_tar_keeping_symlinks: {tinfo}')
-            tarball.extract(
-                member=tinfo,
-                path=target_dir,
-                set_attrs=False,
-            )
+
+            if '..' in tarinfo.name:
+                if TRACE: logger.debug(f'_extract_tar: skipping unsupported {tarinfo} with relative path')
+                continue
+
+            if tarinfo.islnk() or tarinfo.issym():
+                try:
+                    target = tarball._find_link_target(tarinfo)
+                    if not target:
+                        if TRACE:
+                            logger.debug(f'_extract_tar: skipping link with missing target: {tarinfo}')
+                        continue
+
+                except Exception:
+                    import traceback
+                    if TRACE:
+                        logger.debug(f'_extract_tar: skipping link with missing target: {tarinfo}: {traceback.format_exc()}')
+                    continue
+
+            tarinfo.mode = 0o755
+            tarinfo.name = tarinfo.name.lstrip('/')
+            tarball.extract(member=tarinfo, path=target_dir, set_attrs=False,)
+
+
+def extract_tar_with_symlinks(location, target_dir):
+    return extract_tar(location, target_dir, skip_symlinks=False)
 
 
 def lower_keys(mapping):