Handle extracting the same file to the same destination

When a source is added multiple times the content after the 2nd
extraction will be the same and no new files/folders are detected.
So `extract_file` would return the parent folder.
However if there was a clear result after the first extraction the same
result should be returned.

This works trivially until something else is extracted to the same
folder between the 2 extraction operations.

Author: Flamefire

easybuild/tools/filetools.py

@@ -508,7 +508,12 @@ def extract_file(fn, dest, cmd=None, extra_options=None, overwrite=False, forced
     previous_paths = set(_get_paths_purged(abs_dest))
     run_shell_cmd(cmd, in_dry_run=forced, hidden=not trace)
     new_paths = set(_get_paths_purged(abs_dest)) - previous_paths
-    if len(new_paths) == 1:
+    if len(new_paths) == 0:
+        _log.warning(f"No new folder/files found after extraction of {fn} in {abs_dest}, "
+                     f"verify that the same file is not extracted multiple times! "
+                     f"Existing paths: {', '.join(previous_paths)}")
+        base_dir = find_base_dir()  # Fallback
+    elif len(new_paths) == 1:
         new_path = new_paths.pop()
         if os.path.isdir(new_path):
             # note: find_base_dir also changes into the base dir!

test/framework/filetools.py

@@ -2622,9 +2622,18 @@ def test_extract_file(self):
         with self.mocked_stdout_stderr():
             path = ft.extract_file(toy_tarball, extraction_path, change_into_dir=False)
             self.assertTrue(os.path.samefile(path, toy_path))
+            # Extracting the same tarball again detects the same folder
+            path = ft.extract_file(toy_tarball, extraction_path, change_into_dir=False)
+            self.assertTrue(os.path.samefile(path, toy_path))
+
             path = ft.extract_file(bar_tarball, extraction_path, change_into_dir=False)
             self.assertTrue(os.path.samefile(path, os.path.join(extraction_path, 'bar-0.0')))
 
+            # Extracting the same tarball again now can't detect the same folder as there is another one next to it.
+            # Should fall back to the parent dir
+            path = ft.extract_file(toy_tarball, extraction_path, change_into_dir=False)
+            self.assertTrue(os.path.samefile(path, extraction_path))
+
             # Contains just a file
             path = ft.extract_file(file_tarball, extraction_path, change_into_dir=False)
             self.assertTrue(os.path.samefile(path, extraction_path))