90
88
def import_git_blob(texts, mapping, path, name, (base_hexsha, hexsha),
91
89
base_inv, parent_id, revision_id,
92
parent_invs, lookup_object, (base_mode, mode), store_updater,
90
parent_invs, lookup_object, (base_mode, mode), store_updater):
94
91
"""Import a git blob object into a bzr repository.
96
93
:param texts: VersionedFiles to add to
98
95
:param blob: A git blob
99
96
:return: Inventory delta for this file
101
if mapping.is_control_file(path):
103
98
if base_hexsha == hexsha and base_mode == mode:
104
99
# If nothing has changed since the base revision, we're done
106
file_id = lookup_file_id(path)
101
file_id = mapping.generate_file_id(path)
107
102
if stat.S_ISLNK(mode):
108
103
cls = InventoryLink
110
105
cls = InventoryFile
111
106
ie = cls(file_id, name.decode("utf-8"), parent_id)
112
if ie.kind == "file":
113
ie.executable = mode_is_executable(mode)
107
ie.executable = mode_is_executable(mode)
114
108
if base_hexsha == hexsha and mode_kind(base_mode) == mode_kind(mode):
115
109
base_ie = base_inv[base_inv.path2id(path)]
116
110
ie.text_size = base_ie.text_size
117
111
ie.text_sha1 = base_ie.text_sha1
118
if ie.kind == "symlink":
119
ie.symlink_target = base_ie.symlink_target
112
ie.symlink_target = base_ie.symlink_target
120
113
if ie.executable == base_ie.executable:
121
114
ie.revision = base_ie.revision
126
119
if ie.kind == "symlink":
127
120
ie.revision = None
128
121
ie.symlink_target = blob.data
130
125
ie.text_size = sum(imap(len, blob.chunked))
131
126
ie.text_sha1 = osutils.sha_strings(blob.chunked)
132
127
# Check what revision we should store
134
for pinv in parent_invs:
129
for pinv in parent_invs[1:]:
136
131
pie = pinv[file_id]
139
if (pie.text_sha1 == ie.text_sha1 and
140
pie.executable == ie.executable and
141
pie.symlink_target == ie.symlink_target):
134
if pie.text_sha1 == ie.text_sha1 and pie.executable == ie.executable and pie.symlink_target == ie.symlink_target:
142
135
# found a revision in one of the parents to use
143
136
ie.revision = pie.revision
145
parent_key = (file_id, pie.revision)
146
if not parent_key in parent_keys:
147
parent_keys.append(parent_key)
138
parent_keys.append((file_id, pie.revision))
148
139
if ie.revision is None:
149
140
# Need to store a new revision
150
141
ie.revision = revision_id
155
146
chunks = blob.chunked
156
texts.insert_record_stream([
157
ChunkedContentFactory((file_id, ie.revision),
158
tuple(parent_keys), ie.text_sha1, chunks)])
147
texts.insert_record_stream([ChunkedContentFactory((file_id, ie.revision), tuple(parent_keys), ie.text_sha1, chunks)])
160
149
if base_hexsha is not None:
161
old_path = path.decode("utf-8") # Renames are not supported yet
150
old_path = path # Renames are not supported yet
162
151
if stat.S_ISDIR(base_mode):
163
152
invdelta.extend(remove_disappeared_children(base_inv, old_path,
164
153
lookup_object(base_hexsha), [], lookup_object))
167
new_path = path.decode("utf-8")
168
invdelta.append((old_path, new_path, file_id, ie))
156
invdelta.append((old_path, path, file_id, ie))
169
157
if base_hexsha != hexsha:
170
store_updater.add_object(blob, ie, path)
158
store_updater.add_object(blob, ie)
179
167
def import_git_submodule(texts, mapping, path, name, (base_hexsha, hexsha),
180
168
base_inv, parent_id, revision_id, parent_invs, lookup_object,
181
(base_mode, mode), store_updater, lookup_file_id):
169
(base_mode, mode), store_updater):
182
170
if base_hexsha == hexsha and base_mode == mode:
184
file_id = lookup_file_id(path)
172
file_id = mapping.generate_file_id(path)
185
173
ie = TreeReference(file_id, name.decode("utf-8"), parent_id)
186
174
ie.revision = revision_id
187
175
if base_hexsha is None:
191
179
ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)
192
texts.insert_record_stream([
193
ChunkedContentFactory((file_id, ie.revision), (), None, [])])
180
texts.insert_record_stream([ChunkedContentFactory((file_id, ie.revision), (), None, [])])
194
181
invdelta = [(oldpath, path, file_id, ie)]
195
182
return invdelta, {}
198
185
def remove_disappeared_children(base_inv, path, base_tree, existing_children,
200
"""Generate an inventory delta for removed children.
202
:param base_inv: Base inventory against which to generate the
204
:param path: Path to process (unicode)
205
:param base_tree: Git Tree base object
206
:param existing_children: Children that still exist
207
:param lookup_object: Lookup a git object by its SHA1
208
:return: Inventory delta, as list
210
assert type(path) is unicode
212
188
for name, mode, hexsha in base_tree.iteritems():
213
189
if name in existing_children:
215
191
c_path = posixpath.join(path, name.decode("utf-8"))
216
file_id = base_inv.path2id(c_path)
217
assert file_id is not None
218
ret.append((c_path, None, file_id, None))
192
ret.append((c_path, None, base_inv.path2id(c_path), None))
219
193
if stat.S_ISDIR(mode):
220
194
ret.extend(remove_disappeared_children(
221
195
base_inv, c_path, lookup_object(hexsha), [], lookup_object))
225
199
def import_git_tree(texts, mapping, path, name, (base_hexsha, hexsha),
226
200
base_inv, parent_id, revision_id, parent_invs,
227
lookup_object, (base_mode, mode), store_updater,
228
lookup_file_id, allow_submodules=False):
201
lookup_object, (base_mode, mode), store_updater, allow_submodules=False):
229
202
"""Import a git tree object into a bzr repository.
231
204
:param texts: VersionedFiles object to add to
232
:param path: Path in the tree (str)
233
:param name: Name of the tree (str)
205
:param path: Path in the tree
234
206
:param tree: A git tree object
235
207
:param base_inv: Base inventory against which to return inventory delta
236
208
:return: Inventory delta for this subtree
238
assert type(path) is str
239
assert type(name) is str
240
210
if base_hexsha == hexsha and base_mode == mode:
241
211
# If nothing has changed since the base revision, we're done
244
file_id = lookup_file_id(path)
214
file_id = mapping.generate_file_id(path)
245
215
# We just have to hope this is indeed utf-8:
246
216
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
247
217
tree = lookup_object(hexsha)
250
220
old_path = None # Newly appeared here
252
222
base_tree = lookup_object(base_hexsha)
253
old_path = path.decode("utf-8") # Renames aren't supported yet
254
new_path = path.decode("utf-8")
223
old_path = path # Renames aren't supported yet
255
224
if base_tree is None or type(base_tree) is not Tree:
256
225
ie.revision = revision_id
257
invdelta.append((old_path, new_path, ie.file_id, ie))
258
texts.insert_record_stream([
259
ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
226
invdelta.append((old_path, path, ie.file_id, ie))
227
texts.insert_record_stream([ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
260
228
# Remember for next time
261
229
existing_children = set()
273
241
child_base_hexsha = None
274
242
child_base_mode = 0
275
243
if stat.S_ISDIR(child_mode):
276
subinvdelta, grandchildmodes = import_git_tree(texts, mapping,
277
child_path, name, (child_base_hexsha, child_hexsha), base_inv,
278
file_id, revision_id, parent_invs, lookup_object,
279
(child_base_mode, child_mode), store_updater, lookup_file_id,
280
allow_submodules=allow_submodules)
244
subinvdelta, grandchildmodes = import_git_tree(
245
texts, mapping, child_path, name,
246
(child_base_hexsha, child_hexsha),
247
base_inv, file_id, revision_id, parent_invs, lookup_object,
248
(child_base_mode, child_mode), store_updater,
249
allow_submodules=allow_submodules)
281
250
elif S_ISGITLINK(child_mode): # submodule
282
251
if not allow_submodules:
283
252
raise SubmodulesRequireSubtrees()
284
subinvdelta, grandchildmodes = import_git_submodule(texts, mapping,
285
child_path, name, (child_base_hexsha, child_hexsha), base_inv,
286
file_id, revision_id, parent_invs, lookup_object,
287
(child_base_mode, child_mode), store_updater, lookup_file_id)
253
subinvdelta, grandchildmodes = import_git_submodule(
254
texts, mapping, child_path, name,
255
(child_base_hexsha, child_hexsha),
256
base_inv, file_id, revision_id, parent_invs, lookup_object,
257
(child_base_mode, child_mode), store_updater)
289
subinvdelta = import_git_blob(texts, mapping, child_path, name,
290
(child_base_hexsha, child_hexsha), base_inv, file_id,
291
revision_id, parent_invs, lookup_object,
292
(child_base_mode, child_mode), store_updater, lookup_file_id)
259
subinvdelta = import_git_blob(texts, mapping,
260
child_path, name, (child_base_hexsha, child_hexsha),
261
base_inv, file_id, revision_id, parent_invs, lookup_object,
262
(child_base_mode, child_mode), store_updater)
293
263
grandchildmodes = {}
294
264
child_modes.update(grandchildmodes)
295
265
invdelta.extend(subinvdelta)
298
268
child_modes[child_path] = child_mode
299
269
# Remove any children that have disappeared
300
270
if base_tree is not None and type(base_tree) is Tree:
301
invdelta.extend(remove_disappeared_children(base_inv, old_path,
271
invdelta.extend(remove_disappeared_children(base_inv, old_path,
302
272
base_tree, existing_children, lookup_object))
303
store_updater.add_object(tree, ie, path)
273
store_updater.add_object(tree, ie)
304
274
return invdelta, child_modes
307
def verify_commit_reconstruction(target_git_object_retriever, lookup_object,
308
o, rev, ret_tree, parent_trees, mapping, unusual_modes):
309
new_unusual_modes = mapping.export_unusual_file_modes(rev)
310
if new_unusual_modes != unusual_modes:
311
raise AssertionError("unusual modes don't match: %r != %r" % (
312
unusual_modes, new_unusual_modes))
313
# Verify that we can reconstruct the commit properly
314
rec_o = target_git_object_retriever._reconstruct_commit(rev, o.tree, True)
316
raise AssertionError("Reconstructed commit differs: %r != %r" % (
320
for path, obj, ie in _tree_to_objects(ret_tree, parent_trees,
321
target_git_object_retriever._cache.idmap, unusual_modes, mapping.BZR_DUMMY_FILE):
322
old_obj_id = tree_lookup_path(lookup_object, o.tree, path)[1]
324
if obj.id != old_obj_id:
325
diff.append((path, lookup_object(old_obj_id), obj))
326
for (path, old_obj, new_obj) in diff:
327
while (old_obj.type_name == "tree" and
328
new_obj.type_name == "tree" and
329
sorted(old_obj) == sorted(new_obj)):
331
if old_obj[name][0] != new_obj[name][0]:
332
raise AssertionError("Modes for %s differ: %o != %o" %
333
(path, old_obj[name][0], new_obj[name][0]))
334
if old_obj[name][1] != new_obj[name][1]:
335
# Found a differing child, delve deeper
336
path = posixpath.join(path, name)
337
old_obj = lookup_object(old_obj[name][1])
338
new_obj = new_objs[path]
340
raise AssertionError("objects differ for %s: %r != %r" % (path,
344
277
def import_git_commit(repo, mapping, head, lookup_object,
345
278
target_git_object_retriever, trees_cache):
346
279
o = lookup_object(head)
347
rev, roundtrip_revid, verifiers = mapping.import_commit(o,
348
lambda x: target_git_object_retriever.lookup_git_sha(x)[1][0])
280
rev = mapping.import_commit(o)
349
281
# We have to do this here, since we have to walk the tree and
350
282
# we need to make sure to import the blobs / trees with the right
351
283
# path; this may involve adding them more than once.
359
291
base_tree = lookup_object(o.parents[0]).tree
360
292
base_mode = stat.S_IFDIR
361
293
store_updater = target_git_object_retriever._get_updater(rev)
362
fileid_map = mapping.get_fileid_map(lookup_object, o.tree)
294
store_updater.add_object(o, None)
363
295
inv_delta, unusual_modes = import_git_tree(repo.texts,
364
mapping, "", "", (base_tree, o.tree), base_inv,
296
mapping, "", u"", (base_tree, o.tree), base_inv,
365
297
None, rev.revision_id, [p.inventory for p in parent_trees],
366
298
lookup_object, (base_mode, stat.S_IFDIR), store_updater,
367
fileid_map.lookup_file_id,
368
299
allow_submodules=getattr(repo._format, "supports_tree_reference", False))
300
store_updater.finish()
369
301
if unusual_modes != {}:
370
302
for path, mode in unusual_modes.iteritems():
371
303
warn_unusual_mode(rev.foreign_revid, path, mode)
376
308
basis_id = NULL_REVISION
378
310
rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,
379
inv_delta, rev.revision_id, rev.parent_ids, base_inv)
380
# FIXME: Check verifiers
381
testament = StrictTestament3(rev, inv)
382
calculated_verifiers = { "testament3-sha1": testament.as_sha1() }
383
if roundtrip_revid is not None:
384
original_revid = rev.revision_id
385
rev.revision_id = roundtrip_revid
386
if calculated_verifiers != verifiers:
387
trace.mutter("Testament SHA1 %r for %r did not match %r.",
388
calculated_verifiers["testament3-sha1"],
389
rev.revision_id, verifiers["testament3-sha1"])
390
rev.revision_id = original_revid
391
store_updater.add_object(o, calculated_verifiers, None)
392
store_updater.finish()
393
ret_tree = RevisionTree(repo, inv, rev.revision_id)
394
trees_cache.add(ret_tree)
311
inv_delta, rev.revision_id, rev.parent_ids,
313
trees_cache.add(RevisionTree(repo, inv, rev.revision_id))
395
314
repo.add_revision(rev.revision_id, rev)
396
315
if "verify" in debug.debug_flags:
397
verify_commit_reconstruction(target_git_object_retriever,
398
lookup_object, o, rev, ret_tree, parent_trees, mapping,
316
new_unusual_modes = mapping.export_unusual_file_modes(rev)
317
if new_unusual_modes != unusual_modes:
318
raise AssertionError("unusual modes don't match: %r != %r" % (unusual_modes, new_unusual_modes))
319
objs = inventory_to_tree_and_blobs(inv, repo.texts, mapping, unusual_modes)
320
for newsha1, newobj, path in objs:
321
assert path is not None
325
(oldmode, oldsha1) = tree_lookup_path(lookup_object, o.tree, path)
326
if oldsha1 != newsha1:
327
raise AssertionError("%r != %r in %s" % (oldsha1, newsha1, path))
402
330
def import_git_objects(repo, mapping, object_iter,
430
358
if isinstance(o, Commit):
431
rev, roundtrip_revid, verifiers = mapping.import_commit(o,
433
if (repo.has_revision(rev.revision_id) or
434
(roundtrip_revid and repo.has_revision(roundtrip_revid))):
359
rev = mapping.import_commit(o)
360
if repo.has_revision(rev.revision_id):
362
squash_revision(repo, rev)
436
363
graph.append((o.id, o.parents))
437
364
heads.extend([p for p in o.parents if p not in checked])
438
365
elif isinstance(o, Tag):
490
418
"""See InterRepository.copy_content."""
491
419
self.fetch(revision_id, pb, find_ghosts=False)
421
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
422
mapping=None, fetch_spec=None):
423
self.fetch_refs(revision_id=revision_id, pb=pb,
424
find_ghosts=find_ghosts, mapping=mapping, fetch_spec=fetch_spec)
494
427
class InterGitNonGitRepository(InterGitRepository):
495
428
"""Base InterRepository that copies revisions from a Git into a non-Git
498
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
499
"""Fetch objects from a remote server.
501
:param determine_wants: determine_wants callback
502
:param mapping: BzrGitMapping to use
503
:param pb: Optional progress bar
504
:param limit: Maximum number of commits to import.
505
:return: Tuple with pack hint, last imported revision id and remote refs
507
raise NotImplementedError(self.fetch_objects)
509
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
431
def fetch_refs(self, revision_id=None, pb=None, find_ghosts=False,
510
432
mapping=None, fetch_spec=None):
511
433
if mapping is None:
512
434
mapping = self.source.get_mapping()
516
438
interesting_heads = fetch_spec.heads
518
440
interesting_heads = None
519
442
def determine_wants(refs):
520
444
if interesting_heads is None:
521
445
ret = [sha for (ref, sha) in refs.iteritems() if not ref.endswith("^{}")]
523
ret = [self.source.lookup_bzr_revision_id(revid)[0] for revid in interesting_heads if revid not in (None, NULL_REVISION)]
524
return [rev for rev in ret if not self.target.has_revision(self.source.lookup_foreign_revision_id(rev))]
525
(pack_hint, _, remote_refs) = self.fetch_objects(determine_wants, mapping, pb)
447
ret = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in interesting_heads if revid not in (None, NULL_REVISION)]
448
return [rev for rev in ret if not self.target.has_revision(mapping.revision_id_foreign_to_bzr(rev))]
449
(pack_hint, _) = self.fetch_objects(determine_wants, mapping, pb)
526
450
if pack_hint is not None and self.target._format.pack_compresses:
527
451
self.target.pack(hint=pack_hint)
452
if interesting_heads is not None:
453
present_interesting_heads = self.target.has_revisions(interesting_heads)
454
missing_interesting_heads = set(interesting_heads) - present_interesting_heads
455
if missing_interesting_heads:
456
raise AssertionError("Missing interesting heads: %r" % missing_interesting_heads)
531
460
_GIT_PROGRESS_RE = re.compile(r"(.*?): +(\d+)% \((\d+)/(\d+)\)")
574
489
heads = self.get_target_heads()
575
490
graph_walker = store.get_graph_walker(
576
491
[store._lookup_revision_sha1(head) for head in heads])
577
wants_recorder = DetermineWantsRecorder(determine_wants)
494
def record_determine_wants(heads):
495
wants = determine_wants(heads)
496
recorded_wants.extend(wants)
581
501
create_pb = pb = ui.ui_factory.nested_progress_bar()
583
503
objects_iter = self.source.fetch_objects(
584
wants_recorder, graph_walker, store.get_raw,
586
(pack_hint, last_rev) = import_git_objects(self.target, mapping,
587
objects_iter, store, wants_recorder.wants, pb, limit)
588
return (pack_hint, last_rev, wants_recorder.remote_refs)
504
record_determine_wants, graph_walker,
505
store.get_raw, progress)
506
return import_git_objects(self.target, mapping,
507
objects_iter, store, recorded_wants, pb, limit)
591
510
create_pb.finished()
644
562
graphwalker = self.target._git.get_graph_walker()
645
563
if (isinstance(self.source, LocalGitRepository) and
646
564
isinstance(self.target, LocalGitRepository)):
647
refs = self.source._git.fetch(self.target._git, determine_wants,
565
return self.source._git.fetch(self.target._git, determine_wants,
649
return (None, None, refs)
650
567
elif (isinstance(self.source, LocalGitRepository) and
651
568
isinstance(self.target, RemoteGitRepository)):
652
569
raise NotImplementedError
654
571
isinstance(self.target, LocalGitRepository)):
655
572
f, commit = self.target._git.object_store.add_thin_pack()
657
refs = self.source.bzrdir.root_transport.fetch_pack(
658
determine_wants, graphwalker, f.write, progress)
574
refs = self.source._git.fetch_pack(determine_wants,
575
graphwalker, f.write, progress)
660
return (None, None, refs)
665
582
raise AssertionError
667
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
584
def fetch_refs(self, revision_id=None, pb=None, find_ghosts=False,
668
585
mapping=None, fetch_spec=None, branches=None):
669
586
if mapping is None:
670
587
mapping = self.source.get_mapping()