test_object_store.py 15 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397
  1. # test_object_store.py -- tests for object_store.py
  2. # Copyright (C) 2008 Jelmer Vernooij <jelmer@jelmer.uk>
  3. #
  4. # SPDX-License-Identifier: Apache-2.0 OR GPL-2.0-or-later
  5. # Dulwich is dual-licensed under the Apache License, Version 2.0 and the GNU
  6. # General Public License as public by the Free Software Foundation; version 2.0
  7. # or (at your option) any later version. You can redistribute it and/or
  8. # modify it under the terms of either of these two licenses.
  9. #
  10. # Unless required by applicable law or agreed to in writing, software
  11. # distributed under the License is distributed on an "AS IS" BASIS,
  12. # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  13. # See the License for the specific language governing permissions and
  14. # limitations under the License.
  15. #
  16. # You should have received a copy of the licenses; if not, see
  17. # <http://www.gnu.org/licenses/> for a copy of the GNU General Public License
  18. # and <http://www.apache.org/licenses/LICENSE-2.0> for a copy of the Apache
  19. # License, Version 2.0.
  20. #
  21. """Tests for the object store interface."""
  22. from typing import TYPE_CHECKING, Any, Callable
  23. from unittest.mock import patch
  24. from dulwich.index import commit_tree
  25. from dulwich.object_store import (
  26. PackBasedObjectStore,
  27. iter_tree_contents,
  28. peel_sha,
  29. )
  30. from dulwich.objects import (
  31. Blob,
  32. Tree,
  33. TreeEntry,
  34. )
  35. from dulwich.protocol import DEPTH_INFINITE
  36. from .utils import make_object, make_tag
  37. if TYPE_CHECKING:
  38. from dulwich.object_store import BaseObjectStore
  39. testobject = make_object(Blob, data=b"yummy data")
  40. class ObjectStoreTests:
  41. store: "BaseObjectStore"
  42. assertEqual: Callable[[object, object], None]
  43. # For type checker purposes - actual implementation supports both styles
  44. assertRaises: Callable[..., Any]
  45. assertNotIn: Callable[[object, object], None]
  46. assertNotEqual: Callable[[object, object], None]
  47. assertIn: Callable[[object, object], None]
  48. assertTrue: Callable[[bool], None]
  49. assertFalse: Callable[[bool], None]
  50. def test_determine_wants_all(self) -> None:
  51. self.assertEqual(
  52. [b"1" * 40],
  53. self.store.determine_wants_all({b"refs/heads/foo": b"1" * 40}),
  54. )
  55. def test_determine_wants_all_zero(self) -> None:
  56. self.assertEqual(
  57. [], self.store.determine_wants_all({b"refs/heads/foo": b"0" * 40})
  58. )
  59. def test_determine_wants_all_depth(self) -> None:
  60. self.store.add_object(testobject)
  61. refs = {b"refs/heads/foo": testobject.id}
  62. with patch.object(self.store, "_get_depth", return_value=1) as m:
  63. self.assertEqual([], self.store.determine_wants_all(refs, depth=0))
  64. self.assertEqual(
  65. [testobject.id],
  66. self.store.determine_wants_all(refs, depth=DEPTH_INFINITE),
  67. )
  68. m.assert_not_called()
  69. self.assertEqual([], self.store.determine_wants_all(refs, depth=1))
  70. m.assert_called_with(testobject.id)
  71. self.assertEqual(
  72. [testobject.id], self.store.determine_wants_all(refs, depth=2)
  73. )
  74. def test_get_depth(self) -> None:
  75. self.assertEqual(0, self.store._get_depth(testobject.id))
  76. self.store.add_object(testobject)
  77. self.assertEqual(
  78. 1, self.store._get_depth(testobject.id, get_parents=lambda x: [])
  79. )
  80. parent = make_object(Blob, data=b"parent data")
  81. self.store.add_object(parent)
  82. self.assertEqual(
  83. 2,
  84. self.store._get_depth(
  85. testobject.id,
  86. get_parents=lambda x: [parent.id] if x == testobject else [],
  87. ),
  88. )
  89. def test_iter(self) -> None:
  90. self.assertEqual([], list(self.store))
  91. def test_get_nonexistant(self) -> None:
  92. self.assertRaises(KeyError, lambda: self.store[b"a" * 40])
  93. def test_contains_nonexistant(self) -> None:
  94. self.assertNotIn(b"a" * 40, self.store)
  95. def test_add_objects_empty(self) -> None:
  96. self.store.add_objects([])
  97. def test_add_commit(self) -> None:
  98. # TODO: Argh, no way to construct Git commit objects without
  99. # access to a serialized form.
  100. self.store.add_objects([])
  101. def test_store_resilience(self) -> None:
  102. """Test if updating an existing stored object doesn't erase the
  103. object from the store.
  104. """
  105. test_object = make_object(Blob, data=b"data")
  106. self.store.add_object(test_object)
  107. test_object_id = test_object.id
  108. test_object.data = test_object.data + b"update"
  109. stored_test_object = self.store[test_object_id]
  110. self.assertNotEqual(test_object.id, stored_test_object.id)
  111. self.assertEqual(stored_test_object.id, test_object_id)
  112. def test_add_object(self) -> None:
  113. self.store.add_object(testobject)
  114. self.assertEqual({testobject.id}, set(self.store))
  115. self.assertIn(testobject.id, self.store)
  116. r = self.store[testobject.id]
  117. self.assertEqual(r, testobject)
  118. def test_add_objects(self) -> None:
  119. data = [(testobject, "mypath")]
  120. self.store.add_objects(data)
  121. self.assertEqual({testobject.id}, set(self.store))
  122. self.assertIn(testobject.id, self.store)
  123. r = self.store[testobject.id]
  124. self.assertEqual(r, testobject)
  125. def test_tree_changes(self) -> None:
  126. blob_a1 = make_object(Blob, data=b"a1")
  127. blob_a2 = make_object(Blob, data=b"a2")
  128. blob_b = make_object(Blob, data=b"b")
  129. for blob in [blob_a1, blob_a2, blob_b]:
  130. self.store.add_object(blob)
  131. blobs_1 = [(b"a", blob_a1.id, 0o100644), (b"b", blob_b.id, 0o100644)]
  132. tree1_id = commit_tree(self.store, blobs_1)
  133. blobs_2 = [(b"a", blob_a2.id, 0o100644), (b"b", blob_b.id, 0o100644)]
  134. tree2_id = commit_tree(self.store, blobs_2)
  135. change_a = (
  136. (b"a", b"a"),
  137. (0o100644, 0o100644),
  138. (blob_a1.id, blob_a2.id),
  139. )
  140. self.assertEqual([change_a], list(self.store.tree_changes(tree1_id, tree2_id)))
  141. self.assertEqual(
  142. [
  143. change_a,
  144. ((b"b", b"b"), (0o100644, 0o100644), (blob_b.id, blob_b.id)),
  145. ],
  146. list(self.store.tree_changes(tree1_id, tree2_id, want_unchanged=True)),
  147. )
  148. def test_iter_tree_contents(self) -> None:
  149. blob_a = make_object(Blob, data=b"a")
  150. blob_b = make_object(Blob, data=b"b")
  151. blob_c = make_object(Blob, data=b"c")
  152. for blob in [blob_a, blob_b, blob_c]:
  153. self.store.add_object(blob)
  154. blobs = [
  155. (b"a", blob_a.id, 0o100644),
  156. (b"ad/b", blob_b.id, 0o100644),
  157. (b"ad/bd/c", blob_c.id, 0o100755),
  158. (b"ad/c", blob_c.id, 0o100644),
  159. (b"c", blob_c.id, 0o100644),
  160. ]
  161. tree_id = commit_tree(self.store, blobs)
  162. self.assertEqual(
  163. [TreeEntry(p, m, h) for (p, h, m) in blobs],
  164. list(iter_tree_contents(self.store, tree_id)),
  165. )
  166. self.assertEqual([], list(iter_tree_contents(self.store, None)))
  167. def test_iter_tree_contents_include_trees(self) -> None:
  168. blob_a = make_object(Blob, data=b"a")
  169. blob_b = make_object(Blob, data=b"b")
  170. blob_c = make_object(Blob, data=b"c")
  171. for blob in [blob_a, blob_b, blob_c]:
  172. self.store.add_object(blob)
  173. blobs = [
  174. (b"a", blob_a.id, 0o100644),
  175. (b"ad/b", blob_b.id, 0o100644),
  176. (b"ad/bd/c", blob_c.id, 0o100755),
  177. ]
  178. tree_id = commit_tree(self.store, blobs)
  179. tree = self.store[tree_id]
  180. assert isinstance(tree, Tree)
  181. tree_ad = self.store[tree[b"ad"][1]]
  182. assert isinstance(tree_ad, Tree)
  183. tree_bd = self.store[tree_ad[b"bd"][1]]
  184. expected = [
  185. TreeEntry(b"", 0o040000, tree_id),
  186. TreeEntry(b"a", 0o100644, blob_a.id),
  187. TreeEntry(b"ad", 0o040000, tree_ad.id),
  188. TreeEntry(b"ad/b", 0o100644, blob_b.id),
  189. TreeEntry(b"ad/bd", 0o040000, tree_bd.id),
  190. TreeEntry(b"ad/bd/c", 0o100755, blob_c.id),
  191. ]
  192. actual = iter_tree_contents(self.store, tree_id, include_trees=True)
  193. self.assertEqual(expected, list(actual))
  194. def make_tag(self, name, obj):
  195. tag = make_tag(obj, name=name)
  196. self.store.add_object(tag)
  197. return tag
  198. def test_peel_sha(self) -> None:
  199. self.store.add_object(testobject)
  200. tag1 = self.make_tag(b"1", testobject)
  201. tag2 = self.make_tag(b"2", testobject)
  202. tag3 = self.make_tag(b"3", testobject)
  203. for obj in [testobject, tag1, tag2, tag3]:
  204. self.assertEqual((obj, testobject), peel_sha(self.store, obj.id))
  205. def test_get_raw(self) -> None:
  206. self.store.add_object(testobject)
  207. self.assertEqual(
  208. (Blob.type_num, b"yummy data"), self.store.get_raw(testobject.id)
  209. )
  210. def test_close(self) -> None:
  211. # For now, just check that close doesn't barf.
  212. self.store.add_object(testobject)
  213. self.store.close()
  214. def test_iter_prefix(self) -> None:
  215. self.store.add_object(testobject)
  216. self.assertEqual([testobject.id], list(self.store.iter_prefix(testobject.id)))
  217. self.assertEqual(
  218. [testobject.id], list(self.store.iter_prefix(testobject.id[:10]))
  219. )
  220. def test_iterobjects_subset_all_present(self) -> None:
  221. """Test iterating over a subset of objects that all exist."""
  222. blob1 = make_object(Blob, data=b"blob 1 data")
  223. blob2 = make_object(Blob, data=b"blob 2 data")
  224. self.store.add_object(blob1)
  225. self.store.add_object(blob2)
  226. objects = list(self.store.iterobjects_subset([blob1.id, blob2.id]))
  227. self.assertEqual(2, len(objects))
  228. object_ids = set(o.id for o in objects)
  229. self.assertEqual(set([blob1.id, blob2.id]), object_ids)
  230. def test_iterobjects_subset_missing_not_allowed(self) -> None:
  231. """Test iterating with missing objects when not allowed."""
  232. blob1 = make_object(Blob, data=b"blob 1 data")
  233. self.store.add_object(blob1)
  234. missing_sha = b"1" * 40
  235. self.assertRaises(
  236. KeyError,
  237. lambda: list(self.store.iterobjects_subset([blob1.id, missing_sha])),
  238. )
  239. def test_iterobjects_subset_missing_allowed(self) -> None:
  240. """Test iterating with missing objects when allowed."""
  241. blob1 = make_object(Blob, data=b"blob 1 data")
  242. self.store.add_object(blob1)
  243. missing_sha = b"1" * 40
  244. objects = list(
  245. self.store.iterobjects_subset([blob1.id, missing_sha], allow_missing=True)
  246. )
  247. self.assertEqual(1, len(objects))
  248. self.assertEqual(blob1.id, objects[0].id)
  249. def test_iter_prefix_not_found(self) -> None:
  250. self.assertEqual([], list(self.store.iter_prefix(b"1" * 40)))
  251. class PackBasedObjectStoreTests(ObjectStoreTests):
  252. store: PackBasedObjectStore
  253. def tearDown(self) -> None:
  254. for pack in self.store.packs:
  255. pack.close()
  256. def test_empty_packs(self) -> None:
  257. self.assertEqual([], list(self.store.packs))
  258. def test_pack_loose_objects(self) -> None:
  259. b1 = make_object(Blob, data=b"yummy data")
  260. self.store.add_object(b1)
  261. b2 = make_object(Blob, data=b"more yummy data")
  262. self.store.add_object(b2)
  263. b3 = make_object(Blob, data=b"even more yummy data")
  264. b4 = make_object(Blob, data=b"and more yummy data")
  265. self.store.add_objects([(b3, None), (b4, None)])
  266. self.assertEqual({b1.id, b2.id, b3.id, b4.id}, set(self.store))
  267. self.assertEqual(1, len(self.store.packs))
  268. self.assertEqual(2, self.store.pack_loose_objects())
  269. self.assertNotEqual([], list(self.store.packs))
  270. self.assertEqual(0, self.store.pack_loose_objects())
  271. def test_repack(self) -> None:
  272. b1 = make_object(Blob, data=b"yummy data")
  273. self.store.add_object(b1)
  274. b2 = make_object(Blob, data=b"more yummy data")
  275. self.store.add_object(b2)
  276. b3 = make_object(Blob, data=b"even more yummy data")
  277. b4 = make_object(Blob, data=b"and more yummy data")
  278. self.store.add_objects([(b3, None), (b4, None)])
  279. b5 = make_object(Blob, data=b"and more data")
  280. b6 = make_object(Blob, data=b"and some more data")
  281. self.store.add_objects([(b5, None), (b6, None)])
  282. self.assertEqual({b1.id, b2.id, b3.id, b4.id, b5.id, b6.id}, set(self.store))
  283. self.assertEqual(2, len(self.store.packs))
  284. self.assertEqual(6, self.store.repack())
  285. self.assertEqual(1, len(self.store.packs))
  286. self.assertEqual(0, self.store.pack_loose_objects())
  287. def test_repack_existing(self) -> None:
  288. b1 = make_object(Blob, data=b"yummy data")
  289. self.store.add_object(b1)
  290. b2 = make_object(Blob, data=b"more yummy data")
  291. self.store.add_object(b2)
  292. self.store.add_objects([(b1, None), (b2, None)])
  293. self.store.add_objects([(b2, None)])
  294. self.assertEqual({b1.id, b2.id}, set(self.store))
  295. self.assertEqual(2, len(self.store.packs))
  296. self.assertEqual(2, self.store.repack())
  297. self.assertEqual(1, len(self.store.packs))
  298. self.assertEqual(0, self.store.pack_loose_objects())
  299. self.assertEqual({b1.id, b2.id}, set(self.store))
  300. self.assertEqual(1, len(self.store.packs))
  301. self.assertEqual(2, self.store.repack())
  302. self.assertEqual(1, len(self.store.packs))
  303. self.assertEqual(0, self.store.pack_loose_objects())
  304. def test_repack_with_exclude(self) -> None:
  305. """Test repacking while excluding specific objects."""
  306. b1 = make_object(Blob, data=b"yummy data")
  307. self.store.add_object(b1)
  308. b2 = make_object(Blob, data=b"more yummy data")
  309. self.store.add_object(b2)
  310. b3 = make_object(Blob, data=b"even more yummy data")
  311. b4 = make_object(Blob, data=b"and more yummy data")
  312. self.store.add_objects([(b3, None), (b4, None)])
  313. self.assertEqual({b1.id, b2.id, b3.id, b4.id}, set(self.store))
  314. self.assertEqual(1, len(self.store.packs))
  315. # Repack, excluding b2 and b3
  316. excluded = {b2.id, b3.id}
  317. self.assertEqual(2, self.store.repack(exclude=excluded))
  318. # Should have repacked only b1 and b4
  319. self.assertEqual(1, len(self.store.packs))
  320. self.assertIn(b1.id, self.store)
  321. self.assertNotIn(b2.id, self.store)
  322. self.assertNotIn(b3.id, self.store)
  323. self.assertIn(b4.id, self.store)
  324. def test_delete_loose_object(self) -> None:
  325. """Test deleting loose objects."""
  326. b1 = make_object(Blob, data=b"test data")
  327. self.store.add_object(b1)
  328. # Verify it's loose
  329. self.assertTrue(self.store.contains_loose(b1.id))
  330. self.assertIn(b1.id, self.store)
  331. # Delete it
  332. self.store.delete_loose_object(b1.id)
  333. # Verify it's gone
  334. self.assertFalse(self.store.contains_loose(b1.id))
  335. self.assertNotIn(b1.id, self.store)