1
# Copyright (C) 2008-2010 Jelmer Vernooij <jelmer@samba.org>
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from dulwich.objects import (
24
from dulwich.object_store import (
27
from itertools import (
40
from bzrlib.errors import (
44
from bzrlib.inventory import (
51
from bzrlib.repository import (
54
from bzrlib.revision import (
57
from bzrlib.revisiontree import (
60
from bzrlib.testament import (
63
from bzrlib.tsort import (
66
from bzrlib.versionedfile import (
67
ChunkedContentFactory,
70
from bzrlib.plugins.git.mapping import (
76
from bzrlib.plugins.git.object_store import (
81
from bzrlib.plugins.git.refs import extract_tags
82
from bzrlib.plugins.git.remote import (
85
from bzrlib.plugins.git.repository import (
92
def import_git_blob(texts, mapping, path, name, (base_hexsha, hexsha),
93
base_inv, parent_id, revision_id,
94
parent_invs, lookup_object, (base_mode, mode), store_updater,
96
"""Import a git blob object into a bzr repository.
98
:param texts: VersionedFiles to add to
99
:param path: Path in the tree
100
:param blob: A git blob
101
:return: Inventory delta for this file
103
if mapping.is_control_file(path):
105
if base_hexsha == hexsha and base_mode == mode:
106
# If nothing has changed since the base revision, we're done
108
file_id = lookup_file_id(path)
109
if stat.S_ISLNK(mode):
113
ie = cls(file_id, name.decode("utf-8"), parent_id)
114
if ie.kind == "file":
115
ie.executable = mode_is_executable(mode)
116
if base_hexsha == hexsha and mode_kind(base_mode) == mode_kind(mode):
117
base_ie = base_inv[base_inv.path2id(path)]
118
ie.text_size = base_ie.text_size
119
ie.text_sha1 = base_ie.text_sha1
120
if ie.kind == "symlink":
121
ie.symlink_target = base_ie.symlink_target
122
if ie.executable == base_ie.executable:
123
ie.revision = base_ie.revision
125
blob = lookup_object(hexsha)
127
blob = lookup_object(hexsha)
128
if ie.kind == "symlink":
130
ie.symlink_target = blob.data
132
ie.text_size = sum(imap(len, blob.chunked))
133
ie.text_sha1 = osutils.sha_strings(blob.chunked)
134
# Check what revision we should store
136
for pinv in parent_invs:
141
if (pie.text_sha1 == ie.text_sha1 and
142
pie.executable == ie.executable and
143
pie.symlink_target == ie.symlink_target):
144
# found a revision in one of the parents to use
145
ie.revision = pie.revision
147
parent_key = (file_id, pie.revision)
148
if not parent_key in parent_keys:
149
parent_keys.append(parent_key)
150
if ie.revision is None:
151
# Need to store a new revision
152
ie.revision = revision_id
153
assert ie.revision is not None
154
if ie.kind == 'symlink':
157
chunks = blob.chunked
158
texts.insert_record_stream([
159
ChunkedContentFactory((file_id, ie.revision),
160
tuple(parent_keys), ie.text_sha1, chunks)])
162
if base_hexsha is not None:
163
old_path = path.decode("utf-8") # Renames are not supported yet
164
if stat.S_ISDIR(base_mode):
165
invdelta.extend(remove_disappeared_children(base_inv, old_path,
166
lookup_object(base_hexsha), [], lookup_object))
169
new_path = path.decode("utf-8")
170
invdelta.append((old_path, new_path, file_id, ie))
171
if base_hexsha != hexsha:
172
store_updater.add_object(blob, ie, path)
176
class SubmodulesRequireSubtrees(BzrError):
177
_fmt = """The repository you are fetching from contains submodules. To continue, upgrade your Bazaar repository to a format that supports nested trees, such as 'development-subtree'."""
181
def import_git_submodule(texts, mapping, path, name, (base_hexsha, hexsha),
182
base_inv, parent_id, revision_id, parent_invs, lookup_object,
183
(base_mode, mode), store_updater, lookup_file_id):
184
if base_hexsha == hexsha and base_mode == mode:
186
file_id = lookup_file_id(path)
187
ie = TreeReference(file_id, name.decode("utf-8"), parent_id)
188
ie.revision = revision_id
189
if base_hexsha is None:
193
ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)
194
texts.insert_record_stream([
195
ChunkedContentFactory((file_id, ie.revision), (), None, [])])
196
invdelta = [(oldpath, path, file_id, ie)]
200
def remove_disappeared_children(base_inv, path, base_tree, existing_children,
202
"""Generate an inventory delta for removed children.
204
:param base_inv: Base inventory against which to generate the
206
:param path: Path to process (unicode)
207
:param base_tree: Git Tree base object
208
:param existing_children: Children that still exist
209
:param lookup_object: Lookup a git object by its SHA1
210
:return: Inventory delta, as list
212
assert type(path) is unicode
214
for name, mode, hexsha in base_tree.iteritems():
215
if name in existing_children:
217
c_path = posixpath.join(path, name.decode("utf-8"))
218
file_id = base_inv.path2id(c_path)
219
assert file_id is not None
220
ret.append((c_path, None, file_id, None))
221
if stat.S_ISDIR(mode):
222
ret.extend(remove_disappeared_children(
223
base_inv, c_path, lookup_object(hexsha), [], lookup_object))
227
def import_git_tree(texts, mapping, path, name, (base_hexsha, hexsha),
228
base_inv, parent_id, revision_id, parent_invs,
229
lookup_object, (base_mode, mode), store_updater,
230
lookup_file_id, allow_submodules=False):
231
"""Import a git tree object into a bzr repository.
233
:param texts: VersionedFiles object to add to
234
:param path: Path in the tree (str)
235
:param name: Name of the tree (str)
236
:param tree: A git tree object
237
:param base_inv: Base inventory against which to return inventory delta
238
:return: Inventory delta for this subtree
240
assert type(path) is str
241
assert type(name) is str
242
if base_hexsha == hexsha and base_mode == mode:
243
# If nothing has changed since the base revision, we're done
246
file_id = lookup_file_id(path)
247
# We just have to hope this is indeed utf-8:
248
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
249
tree = lookup_object(hexsha)
250
if base_hexsha is None:
252
old_path = None # Newly appeared here
254
base_tree = lookup_object(base_hexsha)
255
old_path = path.decode("utf-8") # Renames aren't supported yet
256
new_path = path.decode("utf-8")
257
if base_tree is None or type(base_tree) is not Tree:
258
ie.revision = revision_id
259
invdelta.append((old_path, new_path, ie.file_id, ie))
260
texts.insert_record_stream([
261
ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
262
# Remember for next time
263
existing_children = set()
265
for name, child_mode, child_hexsha in tree.iteritems():
266
existing_children.add(name)
267
child_path = posixpath.join(path, name)
268
if type(base_tree) is Tree:
270
child_base_mode, child_base_hexsha = base_tree[name]
272
child_base_hexsha = None
275
child_base_hexsha = None
277
if stat.S_ISDIR(child_mode):
278
subinvdelta, grandchildmodes = import_git_tree(texts, mapping,
279
child_path, name, (child_base_hexsha, child_hexsha), base_inv,
280
file_id, revision_id, parent_invs, lookup_object,
281
(child_base_mode, child_mode), store_updater, lookup_file_id,
282
allow_submodules=allow_submodules)
283
elif S_ISGITLINK(child_mode): # submodule
284
if not allow_submodules:
285
raise SubmodulesRequireSubtrees()
286
subinvdelta, grandchildmodes = import_git_submodule(texts, mapping,
287
child_path, name, (child_base_hexsha, child_hexsha), base_inv,
288
file_id, revision_id, parent_invs, lookup_object,
289
(child_base_mode, child_mode), store_updater, lookup_file_id)
291
subinvdelta = import_git_blob(texts, mapping, child_path, name,
292
(child_base_hexsha, child_hexsha), base_inv, file_id,
293
revision_id, parent_invs, lookup_object,
294
(child_base_mode, child_mode), store_updater, lookup_file_id)
296
child_modes.update(grandchildmodes)
297
invdelta.extend(subinvdelta)
298
if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
299
stat.S_IFLNK, DEFAULT_FILE_MODE|0111):
300
child_modes[child_path] = child_mode
301
# Remove any children that have disappeared
302
if base_tree is not None and type(base_tree) is Tree:
303
invdelta.extend(remove_disappeared_children(base_inv, old_path,
304
base_tree, existing_children, lookup_object))
305
store_updater.add_object(tree, ie, path)
306
return invdelta, child_modes
309
def verify_commit_reconstruction(target_git_object_retriever, lookup_object,
310
o, rev, ret_tree, parent_trees, mapping, unusual_modes, verifiers):
311
new_unusual_modes = mapping.export_unusual_file_modes(rev)
312
if new_unusual_modes != unusual_modes:
313
raise AssertionError("unusual modes don't match: %r != %r" % (
314
unusual_modes, new_unusual_modes))
315
# Verify that we can reconstruct the commit properly
316
rec_o = target_git_object_retriever._reconstruct_commit(rev, o.tree, True,
319
raise AssertionError("Reconstructed commit differs: %r != %r" % (
323
for path, obj, ie in _tree_to_objects(ret_tree, parent_trees,
324
target_git_object_retriever._cache.idmap, unusual_modes, mapping.BZR_DUMMY_FILE):
325
old_obj_id = tree_lookup_path(lookup_object, o.tree, path)[1]
327
if obj.id != old_obj_id:
328
diff.append((path, lookup_object(old_obj_id), obj))
329
for (path, old_obj, new_obj) in diff:
330
while (old_obj.type_name == "tree" and
331
new_obj.type_name == "tree" and
332
sorted(old_obj) == sorted(new_obj)):
334
if old_obj[name][0] != new_obj[name][0]:
335
raise AssertionError("Modes for %s differ: %o != %o" %
336
(path, old_obj[name][0], new_obj[name][0]))
337
if old_obj[name][1] != new_obj[name][1]:
338
# Found a differing child, delve deeper
339
path = posixpath.join(path, name)
340
old_obj = lookup_object(old_obj[name][1])
341
new_obj = new_objs[path]
343
raise AssertionError("objects differ for %s: %r != %r" % (path,
347
def import_git_commit(repo, mapping, head, lookup_object,
348
target_git_object_retriever, trees_cache):
349
o = lookup_object(head)
350
# Note that this uses mapping.revision_id_foreign_to_bzr. If the parents
351
# were bzr roundtripped revisions they would be specified in the
353
rev, roundtrip_revid, verifiers = mapping.import_commit(
354
o, mapping.revision_id_foreign_to_bzr)
355
# We have to do this here, since we have to walk the tree and
356
# we need to make sure to import the blobs / trees with the right
357
# path; this may involve adding them more than once.
358
parent_trees = trees_cache.revision_trees(rev.parent_ids)
359
if parent_trees == []:
360
base_inv = Inventory(root_id=None)
364
base_inv = parent_trees[0].inventory
365
base_tree = lookup_object(o.parents[0]).tree
366
base_mode = stat.S_IFDIR
367
store_updater = target_git_object_retriever._get_updater(rev)
368
fileid_map = mapping.get_fileid_map(lookup_object, o.tree)
369
inv_delta, unusual_modes = import_git_tree(repo.texts,
370
mapping, "", "", (base_tree, o.tree), base_inv,
371
None, rev.revision_id, [p.inventory for p in parent_trees],
372
lookup_object, (base_mode, stat.S_IFDIR), store_updater,
373
fileid_map.lookup_file_id,
374
allow_submodules=getattr(repo._format, "supports_tree_reference", False))
375
if unusual_modes != {}:
376
for path, mode in unusual_modes.iteritems():
377
warn_unusual_mode(rev.foreign_revid, path, mode)
378
mapping.import_unusual_file_modes(rev, unusual_modes)
380
basis_id = rev.parent_ids[0]
382
basis_id = NULL_REVISION
384
rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,
385
inv_delta, rev.revision_id, rev.parent_ids, base_inv)
387
testament = StrictTestament3(rev, inv)
388
if roundtrip_revid is not None:
389
calculated_verifiers = { "testament3-sha1": testament.as_sha1() }
390
original_revid = rev.revision_id
391
rev.revision_id = roundtrip_revid
392
if calculated_verifiers != verifiers:
393
trace.mutter("Testament SHA1 %r for %r did not match %r.",
394
calculated_verifiers["testament3-sha1"],
395
rev.revision_id, verifiers["testament3-sha1"])
396
rev.revision_id = original_revid
398
calculated_verifiers = {}
399
store_updater.add_object(o, calculated_verifiers, None)
400
store_updater.finish()
401
ret_tree = RevisionTree(repo, inv, rev.revision_id)
402
trees_cache.add(ret_tree)
403
repo.add_revision(rev.revision_id, rev)
404
if "verify" in debug.debug_flags:
405
verify_commit_reconstruction(target_git_object_retriever,
406
lookup_object, o, rev, ret_tree, parent_trees, mapping,
407
unusual_modes, verifiers)
410
def import_git_objects(repo, mapping, object_iter,
411
target_git_object_retriever, heads, pb=None, limit=None):
412
"""Import a set of git objects into a bzr repository.
414
:param repo: Target Bazaar repository
415
:param mapping: Mapping to use
416
:param object_iter: Iterator over Git objects.
417
:return: Tuple with pack hints and last imported revision id
419
def lookup_object(sha):
421
return object_iter[sha]
423
return target_git_object_retriever[sha]
426
heads = list(set(heads))
427
trees_cache = LRUTreeCache(repo)
428
# Find and convert commit objects
431
pb.update("finding revisions to fetch", len(graph), None)
433
assert isinstance(head, str)
435
o = lookup_object(head)
438
if isinstance(o, Commit):
439
rev, roundtrip_revid, verifiers = mapping.import_commit(o,
440
mapping.revision_id_foreign_to_bzr)
441
if (repo.has_revision(rev.revision_id) or
442
(roundtrip_revid and repo.has_revision(roundtrip_revid))):
444
graph.append((o.id, o.parents))
445
heads.extend([p for p in o.parents if p not in checked])
446
elif isinstance(o, Tag):
447
if o.object[1] not in checked:
448
heads.append(o.object[1])
450
trace.warning("Unable to import head object %r" % o)
453
# Order the revisions
454
# Create the inventory objects
456
revision_ids = topo_sort(graph)
458
if limit is not None:
459
revision_ids = revision_ids[:limit]
461
for offset in range(0, len(revision_ids), batch_size):
462
target_git_object_retriever.start_write_group()
464
repo.start_write_group()
466
for i, head in enumerate(
467
revision_ids[offset:offset+batch_size]):
469
pb.update("fetching revisions", offset+i,
471
import_git_commit(repo, mapping, head, lookup_object,
472
target_git_object_retriever, trees_cache)
475
repo.abort_write_group()
478
hint = repo.commit_write_group()
480
pack_hints.extend(hint)
482
target_git_object_retriever.abort_write_group()
485
target_git_object_retriever.commit_write_group()
486
return pack_hints, last_imported
489
class InterGitRepository(InterRepository):
491
_matching_repo_format = GitRepositoryFormat()
493
def _target_has_shas(self, shas):
494
raise NotImplementedError(self._target_has_shas)
496
def get_determine_wants_heads(self, wants, include_tags=False):
498
def determine_wants(refs):
499
potential = set(wants)
501
potential.update([v[1] or v[0] for v in extract_tags(refs).itervalues()])
502
return list(potential - self._target_has_shas(potential))
503
return determine_wants
505
def determine_wants_all(self, refs):
506
potential = set([sha for (ref, sha) in refs.iteritems() if not ref.endswith("^{}")])
507
return list(potential - self._target_has_shas(potential))
510
def _get_repo_format_to_test():
513
def copy_content(self, revision_id=None, pb=None):
514
"""See InterRepository.copy_content."""
515
self.fetch(revision_id, pb, find_ghosts=False)
518
class InterGitNonGitRepository(InterGitRepository):
519
"""Base InterRepository that copies revisions from a Git into a non-Git
522
def _target_has_shas(self, shas):
523
revids = [self.source.lookup_foreign_revision_id(sha) for sha in shas]
524
return self.target.has_revisions(revids)
526
def get_determine_wants_revids(self, revids, include_tags=False):
528
for revid in set(revids):
529
git_sha, mapping = self.source.lookup_bzr_revision_id(revid)
531
return self.get_determine_wants_heads(wants, include_tags=include_tags)
533
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
534
"""Fetch objects from a remote server.
536
:param determine_wants: determine_wants callback
537
:param mapping: BzrGitMapping to use
538
:param pb: Optional progress bar
539
:param limit: Maximum number of commits to import.
540
:return: Tuple with pack hint, last imported revision id and remote refs
542
raise NotImplementedError(self.fetch_objects)
544
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
545
mapping=None, fetch_spec=None):
547
mapping = self.source.get_mapping()
548
if revision_id is not None:
549
interesting_heads = [revision_id]
550
elif fetch_spec is not None:
551
recipe = fetch_spec.get_recipe()
552
if recipe[0] in ("search", "proxy-search"):
553
interesting_heads = recipe[1]
555
raise AssertionError("Unsupported search result type %s" % recipe[0])
557
interesting_heads = None
559
if interesting_heads is not None:
560
determine_wants = self.get_determine_wants_revids(interesting_heads,
563
determine_wants = self.determine_wants_all
565
(pack_hint, _, remote_refs) = self.fetch_objects(determine_wants,
567
if pack_hint is not None and self.target._format.pack_compresses:
568
self.target.pack(hint=pack_hint)
569
assert isinstance(remote_refs, dict)
573
_GIT_PROGRESS_RE = re.compile(r"(.*?): +(\d+)% \((\d+)/(\d+)\)")
574
def report_git_progress(pb, text):
575
text = text.rstrip("\r\n")
576
g = _GIT_PROGRESS_RE.match(text)
578
(text, pct, current, total) = g.groups()
579
pb.update(text, int(current), int(total))
581
pb.update(text, 0, 0)
584
class DetermineWantsRecorder(object):
586
def __init__(self, actual):
589
self.remote_refs = {}
591
def __call__(self, refs):
592
self.remote_refs = refs
593
self.wants = self.actual(refs)
597
class InterRemoteGitNonGitRepository(InterGitNonGitRepository):
598
"""InterRepository that copies revisions from a remote Git into a non-Git
601
def get_target_heads(self):
602
# FIXME: This should be more efficient
603
all_revs = self.target.all_revision_ids()
604
parent_map = self.target.get_parent_map(all_revs)
606
map(all_parents.update, parent_map.itervalues())
607
return set(all_revs) - all_parents
609
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
610
"""See `InterGitNonGitRepository`."""
612
report_git_progress(pb, text)
613
store = BazaarObjectStore(self.target, mapping)
614
self.target.lock_write()
616
heads = self.get_target_heads()
617
graph_walker = store.get_graph_walker(
618
[store._lookup_revision_sha1(head) for head in heads])
619
wants_recorder = DetermineWantsRecorder(determine_wants)
623
create_pb = pb = ui.ui_factory.nested_progress_bar()
625
objects_iter = self.source.fetch_objects(
626
wants_recorder, graph_walker, store.get_raw,
628
trace.mutter("Importing %d new revisions", len(wants_recorder.wants))
629
(pack_hint, last_rev) = import_git_objects(self.target, mapping,
630
objects_iter, store, wants_recorder.wants, pb, limit)
631
return (pack_hint, last_rev, wants_recorder.remote_refs)
639
def is_compatible(source, target):
640
"""Be compatible with GitRepository."""
641
return (isinstance(source, RemoteGitRepository) and
642
target.supports_rich_root() and
643
not isinstance(target, GitRepository) and
644
target.texts is not None)
647
class InterLocalGitNonGitRepository(InterGitNonGitRepository):
648
"""InterRepository that copies revisions from a local Git into a non-Git
651
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
652
"""See `InterGitNonGitRepository`."""
653
remote_refs = self.source._git.get_refs()
654
wants = determine_wants(remote_refs)
657
create_pb = pb = ui.ui_factory.nested_progress_bar()
658
target_git_object_retriever = BazaarObjectStore(self.target, mapping)
660
self.target.lock_write()
662
(pack_hint, last_rev) = import_git_objects(self.target, mapping,
663
self.source._git.object_store,
664
target_git_object_retriever, wants, pb, limit)
665
return (pack_hint, last_rev, remote_refs)
673
def is_compatible(source, target):
674
"""Be compatible with GitRepository."""
675
return (isinstance(source, LocalGitRepository) and
676
target.supports_rich_root() and
677
not isinstance(target, GitRepository) and
678
target.texts is not None)
681
class InterGitGitRepository(InterGitRepository):
682
"""InterRepository that copies between Git repositories."""
684
def fetch_objects(self, determine_wants, mapping, pb=None):
686
trace.note("git: %s", text)
687
graphwalker = self.target._git.get_graph_walker()
688
if (isinstance(self.source, LocalGitRepository) and
689
isinstance(self.target, LocalGitRepository)):
690
refs = self.source._git.fetch(self.target._git, determine_wants,
692
return (None, None, refs)
693
elif (isinstance(self.source, LocalGitRepository) and
694
isinstance(self.target, RemoteGitRepository)):
695
raise NotImplementedError
696
elif (isinstance(self.source, RemoteGitRepository) and
697
isinstance(self.target, LocalGitRepository)):
698
f, commit = self.target._git.object_store.add_thin_pack()
700
refs = self.source.bzrdir.root_transport.fetch_pack(
701
determine_wants, graphwalker, f.write, progress)
703
return (None, None, refs)
710
def _target_has_shas(self, shas):
711
return set([sha for sha in shas if self.target._git.object_store])
713
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
714
mapping=None, fetch_spec=None, branches=None):
716
mapping = self.source.get_mapping()
718
if revision_id is not None:
719
args = [mapping.revision_id_bzr_to_foreign(revision_id)[0]]
720
elif fetch_spec is not None:
721
recipe = fetch_spec.get_recipe()
722
if recipe[0] in ("search", "proxy-search"):
725
raise AssertionError("Unsupported search result type %s" % recipe[0])
726
args = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in heads]
727
if branches is not None:
728
determine_wants = lambda x: [x[y] for y in branches if not x[y] in r.object_store and x[y] != ZERO_SHA]
729
elif fetch_spec is None and revision_id is None:
730
determine_wants = self.determine_wants_all
732
determine_wants = lambda x: [y for y in args if not y in r.object_store and y != ZERO_SHA]
733
wants_recorder = DetermineWantsRecorder(determine_wants)
734
self.fetch_objects(wants_recorder, mapping)
735
return wants_recorder.remote_refs
738
def is_compatible(source, target):
739
"""Be compatible with GitRepository."""
740
return (isinstance(source, GitRepository) and
741
isinstance(target, GitRepository))
743
def get_determine_wants_revids(self, revids, include_tags=False):
745
for revid in set(revids):
746
git_sha, mapping = self.source.lookup_bzr_revision_id(revid)
748
return self.get_determine_wants_heads(wants, include_tags=include_tags)
added
removed
fetch.py
