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 (
23
from dulwich.object_store import (
26
from itertools import (
39
from bzrlib.errors import (
43
from bzrlib.inventory import (
50
from bzrlib.repository import (
53
from bzrlib.revision import (
56
from bzrlib.revisiontree import (
59
from bzrlib.testament import (
62
from bzrlib.tsort import (
65
from bzrlib.versionedfile import (
66
ChunkedContentFactory,
69
from bzrlib.plugins.git.mapping import (
75
from bzrlib.plugins.git.object_store import (
80
from bzrlib.plugins.git.remote import (
83
from bzrlib.plugins.git.repository import (
90
def import_git_blob(texts, mapping, path, name, (base_hexsha, hexsha),
91
base_inv, parent_id, revision_id,
92
parent_invs, lookup_object, (base_mode, mode), store_updater,
94
"""Import a git blob object into a bzr repository.
96
:param texts: VersionedFiles to add to
97
:param path: Path in the tree
98
:param blob: A git blob
99
:return: Inventory delta for this file
101
if mapping.is_control_file(path):
103
if base_hexsha == hexsha and base_mode == mode:
104
# If nothing has changed since the base revision, we're done
106
file_id = lookup_file_id(path)
107
if stat.S_ISLNK(mode):
111
ie = cls(file_id, name.decode("utf-8"), parent_id)
112
if ie.kind == "file":
113
ie.executable = mode_is_executable(mode)
114
if base_hexsha == hexsha and mode_kind(base_mode) == mode_kind(mode):
115
base_ie = base_inv[base_inv.path2id(path)]
116
ie.text_size = base_ie.text_size
117
ie.text_sha1 = base_ie.text_sha1
118
if ie.kind == "symlink":
119
ie.symlink_target = base_ie.symlink_target
120
if ie.executable == base_ie.executable:
121
ie.revision = base_ie.revision
123
blob = lookup_object(hexsha)
125
blob = lookup_object(hexsha)
126
if ie.kind == "symlink":
128
ie.symlink_target = blob.data
130
ie.text_size = sum(imap(len, blob.chunked))
131
ie.text_sha1 = osutils.sha_strings(blob.chunked)
132
# Check what revision we should store
134
for pinv in parent_invs:
139
if (pie.text_sha1 == ie.text_sha1 and
140
pie.executable == ie.executable and
141
pie.symlink_target == ie.symlink_target):
142
# found a revision in one of the parents to use
143
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)
148
if ie.revision is None:
149
# Need to store a new revision
150
ie.revision = revision_id
151
assert ie.revision is not None
152
if ie.kind == 'symlink':
155
chunks = blob.chunked
156
texts.insert_record_stream([
157
ChunkedContentFactory((file_id, ie.revision),
158
tuple(parent_keys), ie.text_sha1, chunks)])
160
if base_hexsha is not None:
161
old_path = path.decode("utf-8") # Renames are not supported yet
162
if stat.S_ISDIR(base_mode):
163
invdelta.extend(remove_disappeared_children(base_inv, old_path,
164
lookup_object(base_hexsha), [], lookup_object))
167
new_path = path.decode("utf-8")
168
invdelta.append((old_path, new_path, file_id, ie))
169
if base_hexsha != hexsha:
170
store_updater.add_object(blob, ie, path)
174
class SubmodulesRequireSubtrees(BzrError):
175
_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'."""
179
def import_git_submodule(texts, mapping, path, name, (base_hexsha, hexsha),
180
base_inv, parent_id, revision_id, parent_invs, lookup_object,
181
(base_mode, mode), store_updater, lookup_file_id):
182
if base_hexsha == hexsha and base_mode == mode:
184
file_id = lookup_file_id(path)
185
ie = TreeReference(file_id, name.decode("utf-8"), parent_id)
186
ie.revision = revision_id
187
if base_hexsha is None:
191
ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)
192
texts.insert_record_stream([
193
ChunkedContentFactory((file_id, ie.revision), (), None, [])])
194
invdelta = [(oldpath, path, file_id, ie)]
198
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
for name, mode, hexsha in base_tree.iteritems():
213
if name in existing_children:
215
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))
219
if stat.S_ISDIR(mode):
220
ret.extend(remove_disappeared_children(
221
base_inv, c_path, lookup_object(hexsha), [], lookup_object))
225
def import_git_tree(texts, mapping, path, name, (base_hexsha, hexsha),
226
base_inv, parent_id, revision_id, parent_invs,
227
lookup_object, (base_mode, mode), store_updater,
228
lookup_file_id, allow_submodules=False):
229
"""Import a git tree object into a bzr repository.
231
:param texts: VersionedFiles object to add to
232
:param path: Path in the tree (str)
233
:param name: Name of the tree (str)
234
:param tree: A git tree object
235
:param base_inv: Base inventory against which to return inventory delta
236
:return: Inventory delta for this subtree
238
assert type(path) is str
239
assert type(name) is str
240
if base_hexsha == hexsha and base_mode == mode:
241
# If nothing has changed since the base revision, we're done
244
file_id = lookup_file_id(path)
245
# We just have to hope this is indeed utf-8:
246
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
247
tree = lookup_object(hexsha)
248
if base_hexsha is None:
250
old_path = None # Newly appeared here
252
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")
255
if base_tree is None or type(base_tree) is not Tree:
256
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, [])])
260
# Remember for next time
261
existing_children = set()
263
for child_mode, name, child_hexsha in tree.entries():
264
existing_children.add(name)
265
child_path = posixpath.join(path, name)
266
if type(base_tree) is Tree:
268
child_base_mode, child_base_hexsha = base_tree[name]
270
child_base_hexsha = None
273
child_base_hexsha = None
275
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)
281
elif S_ISGITLINK(child_mode): # submodule
282
if not allow_submodules:
283
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)
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)
294
child_modes.update(grandchildmodes)
295
invdelta.extend(subinvdelta)
296
if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
297
stat.S_IFLNK, DEFAULT_FILE_MODE|0111):
298
child_modes[child_path] = child_mode
299
# Remove any children that have disappeared
300
if base_tree is not None and type(base_tree) is Tree:
301
invdelta.extend(remove_disappeared_children(base_inv, old_path,
302
base_tree, existing_children, lookup_object))
303
store_updater.add_object(tree, ie, path)
304
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, verifiers):
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,
317
raise AssertionError("Reconstructed commit differs: %r != %r" % (
321
for path, obj, ie in _tree_to_objects(ret_tree, parent_trees,
322
target_git_object_retriever._cache.idmap, unusual_modes, mapping.BZR_DUMMY_FILE):
323
old_obj_id = tree_lookup_path(lookup_object, o.tree, path)[1]
325
if obj.id != old_obj_id:
326
diff.append((path, lookup_object(old_obj_id), obj))
327
for (path, old_obj, new_obj) in diff:
328
while (old_obj.type_name == "tree" and
329
new_obj.type_name == "tree" and
330
sorted(old_obj) == sorted(new_obj)):
332
if old_obj[name][0] != new_obj[name][0]:
333
raise AssertionError("Modes for %s differ: %o != %o" %
334
(path, old_obj[name][0], new_obj[name][0]))
335
if old_obj[name][1] != new_obj[name][1]:
336
# Found a differing child, delve deeper
337
path = posixpath.join(path, name)
338
old_obj = lookup_object(old_obj[name][1])
339
new_obj = new_objs[path]
341
raise AssertionError("objects differ for %s: %r != %r" % (path,
345
def import_git_commit(repo, mapping, head, lookup_object,
346
target_git_object_retriever, trees_cache):
347
o = lookup_object(head)
348
rev, roundtrip_revid, verifiers = mapping.import_commit(o,
349
lambda x: target_git_object_retriever.lookup_git_sha(x)[1][0])
350
# We have to do this here, since we have to walk the tree and
351
# we need to make sure to import the blobs / trees with the right
352
# path; this may involve adding them more than once.
353
parent_trees = trees_cache.revision_trees(rev.parent_ids)
354
if parent_trees == []:
355
base_inv = Inventory(root_id=None)
359
base_inv = parent_trees[0].inventory
360
base_tree = lookup_object(o.parents[0]).tree
361
base_mode = stat.S_IFDIR
362
store_updater = target_git_object_retriever._get_updater(rev)
363
fileid_map = mapping.get_fileid_map(lookup_object, o.tree)
364
inv_delta, unusual_modes = import_git_tree(repo.texts,
365
mapping, "", "", (base_tree, o.tree), base_inv,
366
None, rev.revision_id, [p.inventory for p in parent_trees],
367
lookup_object, (base_mode, stat.S_IFDIR), store_updater,
368
fileid_map.lookup_file_id,
369
allow_submodules=getattr(repo._format, "supports_tree_reference", False))
370
if unusual_modes != {}:
371
for path, mode in unusual_modes.iteritems():
372
warn_unusual_mode(rev.foreign_revid, path, mode)
373
mapping.import_unusual_file_modes(rev, unusual_modes)
375
basis_id = rev.parent_ids[0]
377
basis_id = NULL_REVISION
379
rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,
380
inv_delta, rev.revision_id, rev.parent_ids, base_inv)
382
testament = StrictTestament3(rev, inv)
383
calculated_verifiers = { "testament3-sha1": testament.as_sha1() }
384
if roundtrip_revid is not None:
385
original_revid = rev.revision_id
386
rev.revision_id = roundtrip_revid
387
if calculated_verifiers != verifiers:
388
trace.mutter("Testament SHA1 %r for %r did not match %r.",
389
calculated_verifiers["testament3-sha1"],
390
rev.revision_id, verifiers["testament3-sha1"])
391
rev.revision_id = original_revid
392
store_updater.add_object(o, calculated_verifiers, None)
393
store_updater.finish()
394
ret_tree = RevisionTree(repo, inv, rev.revision_id)
395
trees_cache.add(ret_tree)
396
repo.add_revision(rev.revision_id, rev)
397
if "verify" in debug.debug_flags:
398
verify_commit_reconstruction(target_git_object_retriever,
399
lookup_object, o, rev, ret_tree, parent_trees, mapping,
400
unusual_modes, verifiers)
403
def import_git_objects(repo, mapping, object_iter,
404
target_git_object_retriever, heads, pb=None, limit=None):
405
"""Import a set of git objects into a bzr repository.
407
:param repo: Target Bazaar repository
408
:param mapping: Mapping to use
409
:param object_iter: Iterator over Git objects.
410
:return: Tuple with pack hints and last imported revision id
412
def lookup_object(sha):
414
return object_iter[sha]
416
return target_git_object_retriever[sha]
419
heads = list(set(heads))
420
trees_cache = LRUTreeCache(repo)
421
# Find and convert commit objects
424
pb.update("finding revisions to fetch", len(graph), None)
426
assert isinstance(head, str)
428
o = lookup_object(head)
431
if isinstance(o, Commit):
432
rev, roundtrip_revid, verifiers = mapping.import_commit(o,
434
if (repo.has_revision(rev.revision_id) or
435
(roundtrip_revid and repo.has_revision(roundtrip_revid))):
437
graph.append((o.id, o.parents))
438
heads.extend([p for p in o.parents if p not in checked])
439
elif isinstance(o, Tag):
440
if o.object[1] not in checked:
441
heads.append(o.object[1])
443
trace.warning("Unable to import head object %r" % o)
446
# Order the revisions
447
# Create the inventory objects
449
revision_ids = topo_sort(graph)
451
if limit is not None:
452
revision_ids = revision_ids[:limit]
454
for offset in range(0, len(revision_ids), batch_size):
455
target_git_object_retriever.start_write_group()
457
repo.start_write_group()
459
for i, head in enumerate(
460
revision_ids[offset:offset+batch_size]):
462
pb.update("fetching revisions", offset+i,
464
import_git_commit(repo, mapping, head, lookup_object,
465
target_git_object_retriever, trees_cache)
468
repo.abort_write_group()
471
hint = repo.commit_write_group()
473
pack_hints.extend(hint)
475
target_git_object_retriever.abort_write_group()
478
target_git_object_retriever.commit_write_group()
479
return pack_hints, last_imported
482
class InterGitRepository(InterRepository):
484
_matching_repo_format = GitRepositoryFormat()
487
def _get_repo_format_to_test():
490
def copy_content(self, revision_id=None, pb=None):
491
"""See InterRepository.copy_content."""
492
self.fetch(revision_id, pb, find_ghosts=False)
495
class InterGitNonGitRepository(InterGitRepository):
496
"""Base InterRepository that copies revisions from a Git into a non-Git
499
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
500
"""Fetch objects from a remote server.
502
:param determine_wants: determine_wants callback
503
:param mapping: BzrGitMapping to use
504
:param pb: Optional progress bar
505
:param limit: Maximum number of commits to import.
506
:return: Tuple with pack hint, last imported revision id and remote refs
508
raise NotImplementedError(self.fetch_objects)
510
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
511
mapping=None, fetch_spec=None):
513
mapping = self.source.get_mapping()
514
if revision_id is not None:
515
interesting_heads = [revision_id]
516
elif fetch_spec is not None:
517
interesting_heads = fetch_spec.heads
519
interesting_heads = None
520
def determine_wants(refs):
521
if interesting_heads is None:
522
ret = [sha for (ref, sha) in refs.iteritems() if not ref.endswith("^{}")]
524
ret = [self.source.lookup_bzr_revision_id(revid)[0] for revid in interesting_heads if revid not in (None, NULL_REVISION)]
525
return [rev for rev in ret if not self.target.has_revision(self.source.lookup_foreign_revision_id(rev))]
526
(pack_hint, _, remote_refs) = self.fetch_objects(determine_wants, mapping, pb)
527
if pack_hint is not None and self.target._format.pack_compresses:
528
self.target.pack(hint=pack_hint)
532
_GIT_PROGRESS_RE = re.compile(r"(.*?): +(\d+)% \((\d+)/(\d+)\)")
533
def report_git_progress(pb, text):
534
text = text.rstrip("\r\n")
535
g = _GIT_PROGRESS_RE.match(text)
537
(text, pct, current, total) = g.groups()
538
pb.update(text, int(current), int(total))
540
pb.update(text, 0, 0)
543
class DetermineWantsRecorder(object):
545
def __init__(self, actual):
548
self.remote_refs = {}
550
def __call__(self, refs):
551
self.remote_refs = refs
552
self.wants = self.actual(refs)
556
class InterRemoteGitNonGitRepository(InterGitNonGitRepository):
557
"""InterRepository that copies revisions from a remote Git into a non-Git
560
def get_target_heads(self):
561
# FIXME: This should be more efficient
562
all_revs = self.target.all_revision_ids()
563
parent_map = self.target.get_parent_map(all_revs)
565
map(all_parents.update, parent_map.itervalues())
566
return set(all_revs) - all_parents
568
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
569
"""See `InterGitNonGitRepository`."""
571
report_git_progress(pb, text)
572
store = BazaarObjectStore(self.target, mapping)
573
self.target.lock_write()
575
heads = self.get_target_heads()
576
graph_walker = store.get_graph_walker(
577
[store._lookup_revision_sha1(head) for head in heads])
578
wants_recorder = DetermineWantsRecorder(determine_wants)
582
create_pb = pb = ui.ui_factory.nested_progress_bar()
584
objects_iter = self.source.fetch_objects(
585
wants_recorder, graph_walker, store.get_raw,
587
(pack_hint, last_rev) = import_git_objects(self.target, mapping,
588
objects_iter, store, wants_recorder.wants, pb, limit)
589
return (pack_hint, last_rev, wants_recorder.remote_refs)
597
def is_compatible(source, target):
598
"""Be compatible with GitRepository."""
599
return (isinstance(source, RemoteGitRepository) and
600
target.supports_rich_root() and
601
not isinstance(target, GitRepository) and
602
target.texts is not None)
605
class InterLocalGitNonGitRepository(InterGitNonGitRepository):
606
"""InterRepository that copies revisions from a local Git into a non-Git
609
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
610
"""See `InterGitNonGitRepository`."""
611
remote_refs = self.source._git.get_refs()
612
wants = determine_wants(remote_refs)
615
create_pb = pb = ui.ui_factory.nested_progress_bar()
616
target_git_object_retriever = BazaarObjectStore(self.target, mapping)
618
self.target.lock_write()
620
(pack_hint, last_rev) = import_git_objects(self.target, mapping,
621
self.source._git.object_store,
622
target_git_object_retriever, wants, pb, limit)
623
return (pack_hint, last_rev, remote_refs)
631
def is_compatible(source, target):
632
"""Be compatible with GitRepository."""
633
return (isinstance(source, LocalGitRepository) and
634
target.supports_rich_root() and
635
not isinstance(target, GitRepository) and
636
target.texts is not None)
639
class InterGitGitRepository(InterGitRepository):
640
"""InterRepository that copies between Git repositories."""
642
def fetch_objects(self, determine_wants, mapping, pb=None):
644
trace.note("git: %s", text)
645
graphwalker = self.target._git.get_graph_walker()
646
if (isinstance(self.source, LocalGitRepository) and
647
isinstance(self.target, LocalGitRepository)):
648
refs = self.source._git.fetch(self.target._git, determine_wants,
650
return (None, None, refs)
651
elif (isinstance(self.source, LocalGitRepository) and
652
isinstance(self.target, RemoteGitRepository)):
653
raise NotImplementedError
654
elif (isinstance(self.source, RemoteGitRepository) and
655
isinstance(self.target, LocalGitRepository)):
656
f, commit = self.target._git.object_store.add_thin_pack()
658
refs = self.source.bzrdir.root_transport.fetch_pack(
659
determine_wants, graphwalker, f.write, progress)
661
return (None, None, refs)
668
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
669
mapping=None, fetch_spec=None, branches=None):
671
mapping = self.source.get_mapping()
673
if revision_id is not None:
674
args = [mapping.revision_id_bzr_to_foreign(revision_id)[0]]
675
elif fetch_spec is not None:
676
args = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in fetch_spec.heads]
677
if branches is not None:
678
determine_wants = lambda x: [x[y] for y in branches if not x[y] in r.object_store]
679
elif fetch_spec is None and revision_id is None:
680
determine_wants = r.object_store.determine_wants_all
682
determine_wants = lambda x: [y for y in args if not y in r.object_store]
683
self.fetch_objects(determine_wants, mapping)
686
def is_compatible(source, target):
687
"""Be compatible with GitRepository."""
688
return (isinstance(source, GitRepository) and
689
isinstance(target, GitRepository))
added
removed
fetch.py
