)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":7847,"name":"Alistair Coles","email":"alistairncoles@gmail.com","username":"acoles"},"change_message_id":"90c8677f61cb7b60f18bea39ef461602e57fabc3","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"ef74a06b_3094de63","updated":"2026-04-21 09:09:04.000000000","message":"Agree, the setUp is creating a fresh testdir for every test, so it shouldn\u0027t be necessary to re-create subdirs. However, the setUp is also populating the testdir with an object. I prefer each test\u0027s starting state to be only the objects which that test expects. I don\u0027t think it is great that tests therefore have to undo what has been done in setUp, but I\u0027d rather have that than get confused by unexpected objects.","commit_id":"ef559431f970c7495f772b3683e8885db21e7609"}],"test/unit/cli/test_relinker.py":[{"author":{"_account_id":7847,"name":"Alistair Coles","email":"alistairncoles@gmail.com","username":"acoles"},"change_message_id":"90c8677f61cb7b60f18bea39ef461602e57fabc3","unresolved":true,"context_lines":[{"line_number":2599,"context_line":"        self.assertEqual(part1, part2)"},{"line_number":2600,"context_line":""},{"line_number":2601,"context_line":"        policy \u003d 0"},{"line_number":2602,"context_line":"        self._recreate_objects_dir(policy)"},{"line_number":2603,"context_line":"        objdir1, fname1, _ \u003d self._create_object(policy, part1, hash1)"},{"line_number":2604,"context_line":"        objdir2, fname2, _ \u003d self._create_object(policy, part2, hash2)"},{"line_number":2605,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"b945dc04_cbb1023a","side":"PARENT","line":2602,"updated":"2026-04-21 09:09:04.000000000","message":"this is removing the object created during setUp, so on master if I look at the debug logger output (which I often do), I see what I\u0027d expect:\n\n```\ntest INFO: [step\u003drelink] 2 hash dirs processed (2 files, 2 linked, 0 removed, 0 errors)\n```\n\nbut with this patch I see:\n```\ntest INFO: [step\u003drelink] 3 hash dirs processed (3 files, 3 linked, 0 removed, 0 errors)\n```\n\nwhich is unnecessarily confusing.\n\nSo, while the _recreate_objects_dir is not strictly necessary for the test to pass as it is currently written, and despite there being no assertions about the logging in the test, I still think it is useful to start with a clean dir and have only the two deliberately created objects.","commit_id":"0ab7285d14deccfa1a9bc0e86a4187536286cec0"},{"author":{"_account_id":7847,"name":"Alistair Coles","email":"alistairncoles@gmail.com","username":"acoles"},"change_message_id":"90c8677f61cb7b60f18bea39ef461602e57fabc3","unresolved":true,"context_lines":[{"line_number":2598,"context_line":"        part2 \u003d utils.get_partition_for_hash(hash2, self.rb.part_power)"},{"line_number":2599,"context_line":"        self.assertEqual(part1, part2)"},{"line_number":2600,"context_line":""},{"line_number":2601,"context_line":"        objdir1, fname1, _ \u003d self._create_object(self.policy, part1, hash1)"},{"line_number":2602,"context_line":"        objdir2, fname2, _ \u003d self._create_object(self.policy, part2, hash2)"},{"line_number":2603,"context_line":""},{"line_number":2604,"context_line":"        self.rb.prepare_increase_partition_power()"}],"source_content_type":"text/x-python","patch_set":1,"id":"df7e6d7d_209c685a","line":2601,"range":{"start_line":2601,"start_character":49,"end_line":2601,"end_character":60},"updated":"2026-04-21 09:09:04.000000000","message":"+2 for this 👍","commit_id":"ef559431f970c7495f772b3683e8885db21e7609"}]}