def test_detect_corrupt_new(test_dataset: DatasetForTests,
integration_test_data: Path):
# type: (Tuple[Collection, str, str, Path]) -> None
"""If a dataset exists but cannot be read handle as corrupt"""
path = uri_to_local_path(test_dataset.uri)
# Write corrupted file.
os.unlink(str(path))
with path.open('w') as f:
f.write('corruption!')
assert path.exists()
# No dataset in index at the corrupt location, so it should be trashed.
_check_sync(
collection=test_dataset.collection,
expected_paths=[test_dataset.uri],
expected_mismatches=[
mm.UnreadableDataset(None, test_dataset.uri)
],
expected_index_result={},
cache_path=integration_test_data,
fix_settings=dict(trash_missing=True, trash_archived=True, update_locations=True)
)
assert not path.exists(), "Corrupt dataset without sibling should be trashed with trash_archived=True"
def test_detect_corrupt_existing(test_dataset: DatasetForTests,
integration_test_data: Path):
# type: (Tuple[Collection, str, str, Path]) -> None
"""If a dataset exists but cannot be read, report as corrupt"""
path = uri_to_local_path(test_dataset.uri)
test_dataset.add_to_index()
assert path.exists()
# Overwrite with corrupted file.
os.unlink(str(path))
with path.open('w') as f:
f.write('corruption!')
assert path.exists()
# Another dataset exists in the same location
_check_sync(
collection=test_dataset.collection,
expected_paths=[test_dataset.uri],
expected_mismatches=[
# We don't know if it's the same dataset
mm.UnreadableDataset(None, test_dataset.uri)
],
# Unmodified index
expected_index_result=freeze_index(test_dataset.collection.index_),
cache_path=integration_test_data,
fix_settings=dict(trash_missing=True, trash_archived=True, update_locations=True)
)
# If a dataset is in the index pointing to the corrupt location, it shouldn't be trashed with trash_archived=True
assert path.exists(), "Corrupt dataset with sibling in index should not be trashed"