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 (
58
from bzrlib.revisiontree import InventoryRevisionTree
59
except ImportError: # bzr < 2.4
60
from bzrlib.revisiontree import RevisionTree as InventoryRevisionTree
61
from bzrlib.testament import (
64
from bzrlib.tsort import (
67
from bzrlib.versionedfile import (
68
ChunkedContentFactory,
71
from bzrlib.plugins.git.mapping import (
77
from bzrlib.plugins.git.object_store import (
82
from bzrlib.plugins.git.refs import extract_tags
83
from bzrlib.plugins.git.remote import (
86
from bzrlib.plugins.git.repository import (
93
def import_git_blob(texts, mapping, path, name, (base_hexsha, hexsha),
94
base_inv, parent_id, revision_id,
95
parent_invs, lookup_object, (base_mode, mode), store_updater,
97
"""Import a git blob object into a bzr repository.
99
:param texts: VersionedFiles to add to
100
:param path: Path in the tree
101
:param blob: A git blob
102
:return: Inventory delta for this file
104
if mapping.is_control_file(path):
106
if base_hexsha == hexsha and base_mode == mode:
107
# If nothing has changed since the base revision, we're done
109
file_id = lookup_file_id(path)
110
if stat.S_ISLNK(mode):
114
ie = cls(file_id, name.decode("utf-8"), parent_id)
115
if ie.kind == "file":
116
ie.executable = mode_is_executable(mode)
117
if base_hexsha == hexsha and mode_kind(base_mode) == mode_kind(mode):
118
base_ie = base_inv[base_inv.path2id(path)]
119
ie.text_size = base_ie.text_size
120
ie.text_sha1 = base_ie.text_sha1
121
if ie.kind == "symlink":
122
ie.symlink_target = base_ie.symlink_target
123
if ie.executable == base_ie.executable:
124
ie.revision = base_ie.revision
126
blob = lookup_object(hexsha)
128
blob = lookup_object(hexsha)
129
if ie.kind == "symlink":
131
ie.symlink_target = blob.data
133
ie.text_size = sum(imap(len, blob.chunked))
134
ie.text_sha1 = osutils.sha_strings(blob.chunked)
135
# Check what revision we should store
137
for pinv in parent_invs:
142
if (pie.text_sha1 == ie.text_sha1 and
143
pie.executable == ie.executable and
144
pie.symlink_target == ie.symlink_target):
145
# found a revision in one of the parents to use
146
ie.revision = pie.revision
148
parent_key = (file_id, pie.revision)
149
if not parent_key in parent_keys:
150
parent_keys.append(parent_key)
151
if ie.revision is None:
152
# Need to store a new revision
153
ie.revision = revision_id
154
assert ie.revision is not None
155
if ie.kind == 'symlink':
158
chunks = blob.chunked
159
texts.insert_record_stream([
160
ChunkedContentFactory((file_id, ie.revision),
161
tuple(parent_keys), ie.text_sha1, chunks)])
163
if base_hexsha is not None:
164
old_path = path.decode("utf-8") # Renames are not supported yet
165
if stat.S_ISDIR(base_mode):
166
invdelta.extend(remove_disappeared_children(base_inv, old_path,
167
lookup_object(base_hexsha), [], lookup_object))
170
new_path = path.decode("utf-8")
171
invdelta.append((old_path, new_path, file_id, ie))
172
if base_hexsha != hexsha:
173
store_updater.add_object(blob, ie, path)
177
class SubmodulesRequireSubtrees(BzrError):
178
_fmt = ("The repository you are fetching from contains submodules. "
179
"To continue, upgrade your Bazaar repository to a format that "
180
"supports nested trees, such as 'development-subtree'.")
184
def import_git_submodule(texts, mapping, path, name, (base_hexsha, hexsha),
185
base_inv, parent_id, revision_id, parent_invs, lookup_object,
186
(base_mode, mode), store_updater, lookup_file_id):
187
"""Import a git submodule."""
188
if base_hexsha == hexsha and base_mode == mode:
190
file_id = lookup_file_id(path)
191
ie = TreeReference(file_id, name.decode("utf-8"), parent_id)
192
ie.revision = revision_id
193
if base_hexsha is None:
197
ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)
198
texts.insert_record_stream([
199
ChunkedContentFactory((file_id, ie.revision), (), None, [])])
200
invdelta = [(oldpath, path, file_id, ie)]
204
def remove_disappeared_children(base_inv, path, base_tree, existing_children,
206
"""Generate an inventory delta for removed children.
208
:param base_inv: Base inventory against which to generate the
210
:param path: Path to process (unicode)
211
:param base_tree: Git Tree base object
212
:param existing_children: Children that still exist
213
:param lookup_object: Lookup a git object by its SHA1
214
:return: Inventory delta, as list
216
assert type(path) is unicode
218
for name, mode, hexsha in base_tree.iteritems():
219
if name in existing_children:
221
c_path = posixpath.join(path, name.decode("utf-8"))
222
file_id = base_inv.path2id(c_path)
223
assert file_id is not None
224
ret.append((c_path, None, file_id, None))
225
if stat.S_ISDIR(mode):
226
ret.extend(remove_disappeared_children(
227
base_inv, c_path, lookup_object(hexsha), [], lookup_object))
231
def import_git_tree(texts, mapping, path, name, (base_hexsha, hexsha),
232
base_inv, parent_id, revision_id, parent_invs,
233
lookup_object, (base_mode, mode), store_updater,
234
lookup_file_id, allow_submodules=False):
235
"""Import a git tree object into a bzr repository.
237
:param texts: VersionedFiles object to add to
238
:param path: Path in the tree (str)
239
:param name: Name of the tree (str)
240
:param tree: A git tree object
241
:param base_inv: Base inventory against which to return inventory delta
242
:return: Inventory delta for this subtree
244
assert type(path) is str
245
assert type(name) is str
246
if base_hexsha == hexsha and base_mode == mode:
247
# If nothing has changed since the base revision, we're done
250
file_id = lookup_file_id(path)
251
# We just have to hope this is indeed utf-8:
252
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
253
tree = lookup_object(hexsha)
254
if base_hexsha is None:
256
old_path = None # Newly appeared here
258
base_tree = lookup_object(base_hexsha)
259
old_path = path.decode("utf-8") # Renames aren't supported yet
260
new_path = path.decode("utf-8")
261
if base_tree is None or type(base_tree) is not Tree:
262
ie.revision = revision_id
263
invdelta.append((old_path, new_path, ie.file_id, ie))
264
texts.insert_record_stream([
265
ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
266
# Remember for next time
267
existing_children = set()
269
for name, child_mode, child_hexsha in tree.iteritems():
270
existing_children.add(name)
271
child_path = posixpath.join(path, name)
272
if type(base_tree) is Tree:
274
child_base_mode, child_base_hexsha = base_tree[name]
276
child_base_hexsha = None
279
child_base_hexsha = None
281
if stat.S_ISDIR(child_mode):
282
subinvdelta, grandchildmodes = import_git_tree(texts, mapping,
283
child_path, name, (child_base_hexsha, child_hexsha), base_inv,
284
file_id, revision_id, parent_invs, lookup_object,
285
(child_base_mode, child_mode), store_updater, lookup_file_id,
286
allow_submodules=allow_submodules)
287
elif S_ISGITLINK(child_mode): # submodule
288
if not allow_submodules:
289
raise SubmodulesRequireSubtrees()
290
subinvdelta, grandchildmodes = import_git_submodule(texts, mapping,
291
child_path, name, (child_base_hexsha, child_hexsha), base_inv,
292
file_id, revision_id, parent_invs, lookup_object,
293
(child_base_mode, child_mode), store_updater, lookup_file_id)
295
subinvdelta = import_git_blob(texts, mapping, child_path, name,
296
(child_base_hexsha, child_hexsha), base_inv, file_id,
297
revision_id, parent_invs, lookup_object,
298
(child_base_mode, child_mode), store_updater, lookup_file_id)
300
child_modes.update(grandchildmodes)
301
invdelta.extend(subinvdelta)
302
if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
303
stat.S_IFLNK, DEFAULT_FILE_MODE|0111):
304
child_modes[child_path] = child_mode
305
# Remove any children that have disappeared
306
if base_tree is not None and type(base_tree) is Tree:
307
invdelta.extend(remove_disappeared_children(base_inv, old_path,
308
base_tree, existing_children, lookup_object))
309
store_updater.add_object(tree, ie, path)
310
return invdelta, child_modes
313
def verify_commit_reconstruction(target_git_object_retriever, lookup_object,
314
o, rev, ret_tree, parent_trees, mapping, unusual_modes, verifiers):
315
new_unusual_modes = mapping.export_unusual_file_modes(rev)
316
if new_unusual_modes != unusual_modes:
317
raise AssertionError("unusual modes don't match: %r != %r" % (
318
unusual_modes, new_unusual_modes))
319
# Verify that we can reconstruct the commit properly
320
rec_o = target_git_object_retriever._reconstruct_commit(rev, o.tree, True,
323
raise AssertionError("Reconstructed commit differs: %r != %r" % (
327
for path, obj, ie in _tree_to_objects(ret_tree, parent_trees,
328
target_git_object_retriever._cache.idmap, unusual_modes,
329
mapping.BZR_DUMMY_FILE):
330
old_obj_id = tree_lookup_path(lookup_object, o.tree, path)[1]
332
if obj.id != old_obj_id:
333
diff.append((path, lookup_object(old_obj_id), obj))
334
for (path, old_obj, new_obj) in diff:
335
while (old_obj.type_name == "tree" and
336
new_obj.type_name == "tree" and
337
sorted(old_obj) == sorted(new_obj)):
339
if old_obj[name][0] != new_obj[name][0]:
340
raise AssertionError("Modes for %s differ: %o != %o" %
341
(path, old_obj[name][0], new_obj[name][0]))
342
if old_obj[name][1] != new_obj[name][1]:
343
# Found a differing child, delve deeper
344
path = posixpath.join(path, name)
345
old_obj = lookup_object(old_obj[name][1])
346
new_obj = new_objs[path]
348
raise AssertionError("objects differ for %s: %r != %r" % (path,
352
def import_git_commit(repo, mapping, head, lookup_object,
353
target_git_object_retriever, trees_cache):
354
o = lookup_object(head)
355
# Note that this uses mapping.revision_id_foreign_to_bzr. If the parents
356
# were bzr roundtripped revisions they would be specified in the
358
rev, roundtrip_revid, verifiers = mapping.import_commit(
359
o, mapping.revision_id_foreign_to_bzr)
360
# We have to do this here, since we have to walk the tree and
361
# we need to make sure to import the blobs / trees with the right
362
# path; this may involve adding them more than once.
363
parent_trees = trees_cache.revision_trees(rev.parent_ids)
364
if parent_trees == []:
365
base_inv = Inventory(root_id=None)
369
base_inv = parent_trees[0].inventory
370
base_tree = lookup_object(o.parents[0]).tree
371
base_mode = stat.S_IFDIR
372
store_updater = target_git_object_retriever._get_updater(rev)
373
tree_supplement = mapping.get_fileid_map(lookup_object, o.tree)
374
inv_delta, unusual_modes = import_git_tree(repo.texts,
375
mapping, "", "", (base_tree, o.tree), base_inv,
376
None, rev.revision_id, [p.inventory for p in parent_trees],
377
lookup_object, (base_mode, stat.S_IFDIR), store_updater,
378
tree_supplement.lookup_file_id,
379
allow_submodules=getattr(repo._format, "supports_tree_reference",
381
if unusual_modes != {}:
382
for path, mode in unusual_modes.iteritems():
383
warn_unusual_mode(rev.foreign_revid, path, mode)
384
mapping.import_unusual_file_modes(rev, unusual_modes)
386
basis_id = rev.parent_ids[0]
388
basis_id = NULL_REVISION
390
rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,
391
inv_delta, rev.revision_id, rev.parent_ids, base_inv)
392
ret_tree = InventoryRevisionTree(repo, inv, rev.revision_id)
393
if roundtrip_revid is not None:
394
original_revid = rev.revision_id
395
rev.revision_id = roundtrip_revid
398
if getattr(StrictTestament3, "from_revision_tree", None):
399
testament = StrictTestament3(rev, ret_tree)
401
testament = StrictTestament3(rev, inv)
402
calculated_verifiers = { "testament3-sha1": testament.as_sha1() }
403
if calculated_verifiers != verifiers:
404
trace.mutter("Testament SHA1 %r for %r did not match %r.",
405
calculated_verifiers["testament3-sha1"],
406
rev.revision_id, verifiers["testament3-sha1"])
407
rev.revision_id = original_revid
409
calculated_verifiers = {}
411
calculated_verifiers = {}
412
store_updater.add_object(o, calculated_verifiers, None)
413
store_updater.finish()
414
trees_cache.add(ret_tree)
415
repo.add_revision(rev.revision_id, rev)
416
if "verify" in debug.debug_flags:
417
verify_commit_reconstruction(target_git_object_retriever,
418
lookup_object, o, rev, ret_tree, parent_trees, mapping,
419
unusual_modes, verifiers)
422
def import_git_objects(repo, mapping, object_iter,
423
target_git_object_retriever, heads, pb=None, limit=None):
424
"""Import a set of git objects into a bzr repository.
426
:param repo: Target Bazaar repository
427
:param mapping: Mapping to use
428
:param object_iter: Iterator over Git objects.
429
:return: Tuple with pack hints and last imported revision id
431
def lookup_object(sha):
433
return object_iter[sha]
435
return target_git_object_retriever[sha]
438
heads = list(set(heads))
439
trees_cache = LRUTreeCache(repo)
440
# Find and convert commit objects
443
pb.update("finding revisions to fetch", len(graph), None)
445
assert isinstance(head, str)
447
o = lookup_object(head)
450
if isinstance(o, Commit):
451
rev, roundtrip_revid, verifiers = mapping.import_commit(o,
452
mapping.revision_id_foreign_to_bzr)
453
if (repo.has_revision(rev.revision_id) or
454
(roundtrip_revid and repo.has_revision(roundtrip_revid))):
456
graph.append((o.id, o.parents))
457
heads.extend([p for p in o.parents if p not in checked])
458
elif isinstance(o, Tag):
459
if o.object[1] not in checked:
460
heads.append(o.object[1])
462
trace.warning("Unable to import head object %r" % o)
465
# Order the revisions
466
# Create the inventory objects
468
revision_ids = topo_sort(graph)
470
if limit is not None:
471
revision_ids = revision_ids[:limit]
473
for offset in range(0, len(revision_ids), batch_size):
474
target_git_object_retriever.start_write_group()
476
repo.start_write_group()
478
for i, head in enumerate(
479
revision_ids[offset:offset+batch_size]):
481
pb.update("fetching revisions", offset+i,
483
import_git_commit(repo, mapping, head, lookup_object,
484
target_git_object_retriever, trees_cache)
487
repo.abort_write_group()
490
hint = repo.commit_write_group()
492
pack_hints.extend(hint)
494
target_git_object_retriever.abort_write_group()
497
target_git_object_retriever.commit_write_group()
498
return pack_hints, last_imported
501
class InterGitRepository(InterRepository):
503
_matching_repo_format = GitRepositoryFormat()
505
def _target_has_shas(self, shas):
506
raise NotImplementedError(self._target_has_shas)
508
def get_determine_wants_heads(self, wants, include_tags=False):
510
def determine_wants(refs):
511
potential = set(wants)
514
[v[1] or v[0] for v in extract_tags(refs).itervalues()])
515
return list(potential - self._target_has_shas(potential))
516
return determine_wants
518
def determine_wants_all(self, refs):
519
potential = set([sha for (ref, sha) in refs.iteritems() if not
520
ref.endswith("^{}")])
521
return list(potential - self._target_has_shas(potential))
524
def _get_repo_format_to_test():
527
def copy_content(self, revision_id=None, pb=None):
528
"""See InterRepository.copy_content."""
529
self.fetch(revision_id, pb, find_ghosts=False)
532
class InterGitNonGitRepository(InterGitRepository):
533
"""Base InterRepository that copies revisions from a Git into a non-Git
536
def _target_has_shas(self, shas):
537
revids = [self.source.lookup_foreign_revision_id(sha) for sha in shas]
538
return self.target.has_revisions(revids)
540
def get_determine_wants_revids(self, revids, include_tags=False):
542
for revid in set(revids):
543
git_sha, mapping = self.source.lookup_bzr_revision_id(revid)
545
return self.get_determine_wants_heads(wants,
546
include_tags=include_tags)
548
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
549
"""Fetch objects from a remote server.
551
:param determine_wants: determine_wants callback
552
:param mapping: BzrGitMapping to use
553
:param pb: Optional progress bar
554
:param limit: Maximum number of commits to import.
555
:return: Tuple with pack hint, last imported revision id and remote refs
557
raise NotImplementedError(self.fetch_objects)
559
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
560
mapping=None, fetch_spec=None):
562
mapping = self.source.get_mapping()
563
if revision_id is not None:
564
interesting_heads = [revision_id]
565
elif fetch_spec is not None:
566
recipe = fetch_spec.get_recipe()
567
if recipe[0] in ("search", "proxy-search"):
568
interesting_heads = recipe[1]
570
raise AssertionError("Unsupported search result type %s" %
573
interesting_heads = None
575
if interesting_heads is not None:
576
determine_wants = self.get_determine_wants_revids(
577
interesting_heads, include_tags=False)
579
determine_wants = self.determine_wants_all
581
(pack_hint, _, remote_refs) = self.fetch_objects(determine_wants,
583
if pack_hint is not None and self.target._format.pack_compresses:
584
self.target.pack(hint=pack_hint)
585
assert isinstance(remote_refs, dict)
589
_GIT_PROGRESS_RE = re.compile(r"(.*?): +(\d+)% \((\d+)/(\d+)\)")
590
def report_git_progress(pb, text):
591
text = text.rstrip("\r\n")
592
g = _GIT_PROGRESS_RE.match(text)
594
(text, pct, current, total) = g.groups()
595
pb.update(text, int(current), int(total))
597
pb.update(text, 0, 0)
600
class DetermineWantsRecorder(object):
602
def __init__(self, actual):
605
self.remote_refs = {}
607
def __call__(self, refs):
608
self.remote_refs = refs
609
self.wants = self.actual(refs)
613
class InterRemoteGitNonGitRepository(InterGitNonGitRepository):
614
"""InterRepository that copies revisions from a remote Git into a non-Git
617
def get_target_heads(self):
618
# FIXME: This should be more efficient
619
all_revs = self.target.all_revision_ids()
620
parent_map = self.target.get_parent_map(all_revs)
622
map(all_parents.update, parent_map.itervalues())
623
return set(all_revs) - all_parents
625
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
626
"""See `InterGitNonGitRepository`."""
628
report_git_progress(pb, text)
629
store = BazaarObjectStore(self.target, mapping)
632
heads = self.get_target_heads()
633
graph_walker = store.get_graph_walker(
634
[store._lookup_revision_sha1(head) for head in heads])
635
wants_recorder = DetermineWantsRecorder(determine_wants)
639
create_pb = pb = ui.ui_factory.nested_progress_bar()
641
objects_iter = self.source.fetch_objects(
642
wants_recorder, graph_walker, store.get_raw,
644
trace.mutter("Importing %d new revisions",
645
len(wants_recorder.wants))
646
(pack_hint, last_rev) = import_git_objects(self.target,
647
mapping, objects_iter, store, wants_recorder.wants, pb,
649
return (pack_hint, last_rev, wants_recorder.remote_refs)
657
def is_compatible(source, target):
658
"""Be compatible with GitRepository."""
659
if not isinstance(source, RemoteGitRepository):
661
if not target.supports_rich_root():
663
if isinstance(target, GitRepository):
665
if not getattr(target._format, "supports_full_versioned_files", True):
670
class InterLocalGitNonGitRepository(InterGitNonGitRepository):
671
"""InterRepository that copies revisions from a local Git into a non-Git
674
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
675
"""See `InterGitNonGitRepository`."""
676
remote_refs = self.source._git.get_refs()
677
wants = determine_wants(remote_refs)
680
create_pb = pb = ui.ui_factory.nested_progress_bar()
681
target_git_object_retriever = BazaarObjectStore(self.target, mapping)
683
target_git_object_retriever.lock_write()
685
(pack_hint, last_rev) = import_git_objects(self.target,
686
mapping, self.source._git.object_store,
687
target_git_object_retriever, wants, pb, limit)
688
return (pack_hint, last_rev, remote_refs)
690
target_git_object_retriever.unlock()
696
def is_compatible(source, target):
697
"""Be compatible with GitRepository."""
698
if not isinstance(source, LocalGitRepository):
700
if not target.supports_rich_root():
702
if isinstance(target, GitRepository):
704
if not getattr(target._format, "supports_full_versioned_files", True):
709
class InterGitGitRepository(InterGitRepository):
710
"""InterRepository that copies between Git repositories."""
712
def fetch_objects(self, determine_wants, mapping, pb=None):
714
trace.note("git: %s", text)
715
graphwalker = self.target._git.get_graph_walker()
716
if (isinstance(self.source, LocalGitRepository) and
717
isinstance(self.target, LocalGitRepository)):
718
refs = self.source._git.fetch(self.target._git, determine_wants,
720
return (None, None, refs)
721
elif (isinstance(self.source, LocalGitRepository) and
722
isinstance(self.target, RemoteGitRepository)):
723
raise NotImplementedError
724
elif (isinstance(self.source, RemoteGitRepository) and
725
isinstance(self.target, LocalGitRepository)):
726
f, commit = self.target._git.object_store.add_thin_pack()
728
refs = self.source.bzrdir.root_transport.fetch_pack(
729
determine_wants, graphwalker, f.write, progress)
731
return (None, None, refs)
738
def _target_has_shas(self, shas):
739
return set([sha for sha in shas if self.target._git.object_store])
741
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
742
mapping=None, fetch_spec=None, branches=None):
744
mapping = self.source.get_mapping()
746
if revision_id is not None:
747
args = [self.source.lookup_bzr_revision_id(revision_id)[0]]
748
elif fetch_spec is not None:
749
recipe = fetch_spec.get_recipe()
750
if recipe[0] in ("search", "proxy-search"):
753
raise AssertionError(
754
"Unsupported search result type %s" % recipe[0])
755
args = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in
757
if branches is not None:
758
determine_wants = lambda x: [x[y] for y in branches if not x[y] in r.object_store and x[y] != ZERO_SHA]
759
elif fetch_spec is None and revision_id is None:
760
determine_wants = self.determine_wants_all
762
determine_wants = lambda x: [y for y in args if not y in r.object_store and y != ZERO_SHA]
763
wants_recorder = DetermineWantsRecorder(determine_wants)
764
self.fetch_objects(wants_recorder, mapping)
765
return wants_recorder.remote_refs
768
def is_compatible(source, target):
769
"""Be compatible with GitRepository."""
770
return (isinstance(source, GitRepository) and
771
isinstance(target, GitRepository))
773
def get_determine_wants_revids(self, revids, include_tags=False):
775
for revid in set(revids):
776
git_sha, mapping = self.source.lookup_bzr_revision_id(revid)
778
return self.get_determine_wants_heads(wants,
779
include_tags=include_tags)
added
removed
fetch.py
