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. To continue, upgrade your Bazaar repository to a format that supports nested trees, such as 'development-subtree'."""
182
def import_git_submodule(texts, mapping, path, name, (base_hexsha, hexsha),
183
base_inv, parent_id, revision_id, parent_invs, lookup_object,
184
(base_mode, mode), store_updater, lookup_file_id):
185
if base_hexsha == hexsha and base_mode == mode:
187
file_id = lookup_file_id(path)
188
ie = TreeReference(file_id, name.decode("utf-8"), parent_id)
189
ie.revision = revision_id
190
if base_hexsha is None:
194
ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)
195
texts.insert_record_stream([
196
ChunkedContentFactory((file_id, ie.revision), (), None, [])])
197
invdelta = [(oldpath, path, file_id, ie)]
201
def remove_disappeared_children(base_inv, path, base_tree, existing_children,
203
"""Generate an inventory delta for removed children.
205
:param base_inv: Base inventory against which to generate the
207
:param path: Path to process (unicode)
208
:param base_tree: Git Tree base object
209
:param existing_children: Children that still exist
210
:param lookup_object: Lookup a git object by its SHA1
211
:return: Inventory delta, as list
213
assert type(path) is unicode
215
for name, mode, hexsha in base_tree.iteritems():
216
if name in existing_children:
218
c_path = posixpath.join(path, name.decode("utf-8"))
219
file_id = base_inv.path2id(c_path)
220
assert file_id is not None
221
ret.append((c_path, None, file_id, None))
222
if stat.S_ISDIR(mode):
223
ret.extend(remove_disappeared_children(
224
base_inv, c_path, lookup_object(hexsha), [], lookup_object))
228
def import_git_tree(texts, mapping, path, name, (base_hexsha, hexsha),
229
base_inv, parent_id, revision_id, parent_invs,
230
lookup_object, (base_mode, mode), store_updater,
231
lookup_file_id, allow_submodules=False):
232
"""Import a git tree object into a bzr repository.
234
:param texts: VersionedFiles object to add to
235
:param path: Path in the tree (str)
236
:param name: Name of the tree (str)
237
:param tree: A git tree object
238
:param base_inv: Base inventory against which to return inventory delta
239
:return: Inventory delta for this subtree
241
assert type(path) is str
242
assert type(name) is str
243
if base_hexsha == hexsha and base_mode == mode:
244
# If nothing has changed since the base revision, we're done
247
file_id = lookup_file_id(path)
248
# We just have to hope this is indeed utf-8:
249
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
250
tree = lookup_object(hexsha)
251
if base_hexsha is None:
253
old_path = None # Newly appeared here
255
base_tree = lookup_object(base_hexsha)
256
old_path = path.decode("utf-8") # Renames aren't supported yet
257
new_path = path.decode("utf-8")
258
if base_tree is None or type(base_tree) is not Tree:
259
ie.revision = revision_id
260
invdelta.append((old_path, new_path, ie.file_id, ie))
261
texts.insert_record_stream([
262
ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
263
# Remember for next time
264
existing_children = set()
266
for name, child_mode, child_hexsha in tree.iteritems():
267
existing_children.add(name)
268
child_path = posixpath.join(path, name)
269
if type(base_tree) is Tree:
271
child_base_mode, child_base_hexsha = base_tree[name]
273
child_base_hexsha = None
276
child_base_hexsha = None
278
if stat.S_ISDIR(child_mode):
279
subinvdelta, grandchildmodes = import_git_tree(texts, mapping,
280
child_path, name, (child_base_hexsha, child_hexsha), base_inv,
281
file_id, revision_id, parent_invs, lookup_object,
282
(child_base_mode, child_mode), store_updater, lookup_file_id,
283
allow_submodules=allow_submodules)
284
elif S_ISGITLINK(child_mode): # submodule
285
if not allow_submodules:
286
raise SubmodulesRequireSubtrees()
287
subinvdelta, grandchildmodes = import_git_submodule(texts, mapping,
288
child_path, name, (child_base_hexsha, child_hexsha), base_inv,
289
file_id, revision_id, parent_invs, lookup_object,
290
(child_base_mode, child_mode), store_updater, lookup_file_id)
292
subinvdelta = import_git_blob(texts, mapping, child_path, name,
293
(child_base_hexsha, child_hexsha), base_inv, file_id,
294
revision_id, parent_invs, lookup_object,
295
(child_base_mode, child_mode), store_updater, lookup_file_id)
297
child_modes.update(grandchildmodes)
298
invdelta.extend(subinvdelta)
299
if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
300
stat.S_IFLNK, DEFAULT_FILE_MODE|0111):
301
child_modes[child_path] = child_mode
302
# Remove any children that have disappeared
303
if base_tree is not None and type(base_tree) is Tree:
304
invdelta.extend(remove_disappeared_children(base_inv, old_path,
305
base_tree, existing_children, lookup_object))
306
store_updater.add_object(tree, ie, path)
307
return invdelta, child_modes
310
def verify_commit_reconstruction(target_git_object_retriever, lookup_object,
311
o, rev, ret_tree, parent_trees, mapping, unusual_modes, verifiers):
312
new_unusual_modes = mapping.export_unusual_file_modes(rev)
313
if new_unusual_modes != unusual_modes:
314
raise AssertionError("unusual modes don't match: %r != %r" % (
315
unusual_modes, new_unusual_modes))
316
# Verify that we can reconstruct the commit properly
317
rec_o = target_git_object_retriever._reconstruct_commit(rev, o.tree, True,
320
raise AssertionError("Reconstructed commit differs: %r != %r" % (
324
for path, obj, ie in _tree_to_objects(ret_tree, parent_trees,
325
target_git_object_retriever._cache.idmap, unusual_modes, mapping.BZR_DUMMY_FILE):
326
old_obj_id = tree_lookup_path(lookup_object, o.tree, path)[1]
328
if obj.id != old_obj_id:
329
diff.append((path, lookup_object(old_obj_id), obj))
330
for (path, old_obj, new_obj) in diff:
331
while (old_obj.type_name == "tree" and
332
new_obj.type_name == "tree" and
333
sorted(old_obj) == sorted(new_obj)):
335
if old_obj[name][0] != new_obj[name][0]:
336
raise AssertionError("Modes for %s differ: %o != %o" %
337
(path, old_obj[name][0], new_obj[name][0]))
338
if old_obj[name][1] != new_obj[name][1]:
339
# Found a differing child, delve deeper
340
path = posixpath.join(path, name)
341
old_obj = lookup_object(old_obj[name][1])
342
new_obj = new_objs[path]
344
raise AssertionError("objects differ for %s: %r != %r" % (path,
348
def import_git_commit(repo, mapping, head, lookup_object,
349
target_git_object_retriever, trees_cache):
350
o = lookup_object(head)
351
# Note that this uses mapping.revision_id_foreign_to_bzr. If the parents
352
# were bzr roundtripped revisions they would be specified in the
354
rev, roundtrip_revid, verifiers = mapping.import_commit(
355
o, mapping.revision_id_foreign_to_bzr)
356
# We have to do this here, since we have to walk the tree and
357
# we need to make sure to import the blobs / trees with the right
358
# path; this may involve adding them more than once.
359
parent_trees = trees_cache.revision_trees(rev.parent_ids)
360
if parent_trees == []:
361
base_inv = Inventory(root_id=None)
365
base_inv = parent_trees[0].inventory
366
base_tree = lookup_object(o.parents[0]).tree
367
base_mode = stat.S_IFDIR
368
store_updater = target_git_object_retriever._get_updater(rev)
369
fileid_map = mapping.get_fileid_map(lookup_object, o.tree)
370
inv_delta, unusual_modes = import_git_tree(repo.texts,
371
mapping, "", "", (base_tree, o.tree), base_inv,
372
None, rev.revision_id, [p.inventory for p in parent_trees],
373
lookup_object, (base_mode, stat.S_IFDIR), store_updater,
374
fileid_map.lookup_file_id,
375
allow_submodules=getattr(repo._format, "supports_tree_reference", False))
376
if unusual_modes != {}:
377
for path, mode in unusual_modes.iteritems():
378
warn_unusual_mode(rev.foreign_revid, path, mode)
379
mapping.import_unusual_file_modes(rev, unusual_modes)
381
basis_id = rev.parent_ids[0]
383
basis_id = NULL_REVISION
385
rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,
386
inv_delta, rev.revision_id, rev.parent_ids, base_inv)
388
testament = StrictTestament3(rev, inv)
389
if roundtrip_revid is not None:
390
calculated_verifiers = { "testament3-sha1": testament.as_sha1() }
391
original_revid = rev.revision_id
392
rev.revision_id = roundtrip_revid
393
if calculated_verifiers != verifiers:
394
trace.mutter("Testament SHA1 %r for %r did not match %r.",
395
calculated_verifiers["testament3-sha1"],
396
rev.revision_id, verifiers["testament3-sha1"])
397
rev.revision_id = original_revid
399
calculated_verifiers = {}
400
store_updater.add_object(o, calculated_verifiers, None)
401
store_updater.finish()
402
ret_tree = InventoryRevisionTree(repo, inv, rev.revision_id)
403
trees_cache.add(ret_tree)
404
repo.add_revision(rev.revision_id, rev)
405
if "verify" in debug.debug_flags:
406
verify_commit_reconstruction(target_git_object_retriever,
407
lookup_object, o, rev, ret_tree, parent_trees, mapping,
408
unusual_modes, verifiers)
411
def import_git_objects(repo, mapping, object_iter,
412
target_git_object_retriever, heads, pb=None, limit=None):
413
"""Import a set of git objects into a bzr repository.
415
:param repo: Target Bazaar repository
416
:param mapping: Mapping to use
417
:param object_iter: Iterator over Git objects.
418
:return: Tuple with pack hints and last imported revision id
420
def lookup_object(sha):
422
return object_iter[sha]
424
return target_git_object_retriever[sha]
427
heads = list(set(heads))
428
trees_cache = LRUTreeCache(repo)
429
# Find and convert commit objects
432
pb.update("finding revisions to fetch", len(graph), None)
434
assert isinstance(head, str)
436
o = lookup_object(head)
439
if isinstance(o, Commit):
440
rev, roundtrip_revid, verifiers = mapping.import_commit(o,
441
mapping.revision_id_foreign_to_bzr)
442
if (repo.has_revision(rev.revision_id) or
443
(roundtrip_revid and repo.has_revision(roundtrip_revid))):
445
graph.append((o.id, o.parents))
446
heads.extend([p for p in o.parents if p not in checked])
447
elif isinstance(o, Tag):
448
if o.object[1] not in checked:
449
heads.append(o.object[1])
451
trace.warning("Unable to import head object %r" % o)
454
# Order the revisions
455
# Create the inventory objects
457
revision_ids = topo_sort(graph)
459
if limit is not None:
460
revision_ids = revision_ids[:limit]
462
for offset in range(0, len(revision_ids), batch_size):
463
target_git_object_retriever.start_write_group()
465
repo.start_write_group()
467
for i, head in enumerate(
468
revision_ids[offset:offset+batch_size]):
470
pb.update("fetching revisions", offset+i,
472
import_git_commit(repo, mapping, head, lookup_object,
473
target_git_object_retriever, trees_cache)
476
repo.abort_write_group()
479
hint = repo.commit_write_group()
481
pack_hints.extend(hint)
483
target_git_object_retriever.abort_write_group()
486
target_git_object_retriever.commit_write_group()
487
return pack_hints, last_imported
490
class InterGitRepository(InterRepository):
492
_matching_repo_format = GitRepositoryFormat()
494
def _target_has_shas(self, shas):
495
raise NotImplementedError(self._target_has_shas)
497
def get_determine_wants_heads(self, wants, include_tags=False):
499
def determine_wants(refs):
500
potential = set(wants)
502
potential.update([v[1] or v[0] for v in extract_tags(refs).itervalues()])
503
return list(potential - self._target_has_shas(potential))
504
return determine_wants
506
def determine_wants_all(self, refs):
507
potential = set([sha for (ref, sha) in refs.iteritems() if not ref.endswith("^{}")])
508
return list(potential - self._target_has_shas(potential))
511
def _get_repo_format_to_test():
514
def copy_content(self, revision_id=None, pb=None):
515
"""See InterRepository.copy_content."""
516
self.fetch(revision_id, pb, find_ghosts=False)
519
class InterGitNonGitRepository(InterGitRepository):
520
"""Base InterRepository that copies revisions from a Git into a non-Git
523
def _target_has_shas(self, shas):
524
revids = [self.source.lookup_foreign_revision_id(sha) for sha in shas]
525
return self.target.has_revisions(revids)
527
def get_determine_wants_revids(self, revids, include_tags=False):
529
for revid in set(revids):
530
git_sha, mapping = self.source.lookup_bzr_revision_id(revid)
532
return self.get_determine_wants_heads(wants, include_tags=include_tags)
534
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
535
"""Fetch objects from a remote server.
537
:param determine_wants: determine_wants callback
538
:param mapping: BzrGitMapping to use
539
:param pb: Optional progress bar
540
:param limit: Maximum number of commits to import.
541
:return: Tuple with pack hint, last imported revision id and remote refs
543
raise NotImplementedError(self.fetch_objects)
545
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
546
mapping=None, fetch_spec=None):
548
mapping = self.source.get_mapping()
549
if revision_id is not None:
550
interesting_heads = [revision_id]
551
elif fetch_spec is not None:
552
recipe = fetch_spec.get_recipe()
553
if recipe[0] in ("search", "proxy-search"):
554
interesting_heads = recipe[1]
556
raise AssertionError("Unsupported search result type %s" % recipe[0])
558
interesting_heads = None
560
if interesting_heads is not None:
561
determine_wants = self.get_determine_wants_revids(interesting_heads,
564
determine_wants = self.determine_wants_all
566
(pack_hint, _, remote_refs) = self.fetch_objects(determine_wants,
568
if pack_hint is not None and self.target._format.pack_compresses:
569
self.target.pack(hint=pack_hint)
570
assert isinstance(remote_refs, dict)
574
_GIT_PROGRESS_RE = re.compile(r"(.*?): +(\d+)% \((\d+)/(\d+)\)")
575
def report_git_progress(pb, text):
576
text = text.rstrip("\r\n")
577
g = _GIT_PROGRESS_RE.match(text)
579
(text, pct, current, total) = g.groups()
580
pb.update(text, int(current), int(total))
582
pb.update(text, 0, 0)
585
class DetermineWantsRecorder(object):
587
def __init__(self, actual):
590
self.remote_refs = {}
592
def __call__(self, refs):
593
self.remote_refs = refs
594
self.wants = self.actual(refs)
598
class InterRemoteGitNonGitRepository(InterGitNonGitRepository):
599
"""InterRepository that copies revisions from a remote Git into a non-Git
602
def get_target_heads(self):
603
# FIXME: This should be more efficient
604
all_revs = self.target.all_revision_ids()
605
parent_map = self.target.get_parent_map(all_revs)
607
map(all_parents.update, parent_map.itervalues())
608
return set(all_revs) - all_parents
610
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
611
"""See `InterGitNonGitRepository`."""
613
report_git_progress(pb, text)
614
store = BazaarObjectStore(self.target, mapping)
615
self.target.lock_write()
617
heads = self.get_target_heads()
618
graph_walker = store.get_graph_walker(
619
[store._lookup_revision_sha1(head) for head in heads])
620
wants_recorder = DetermineWantsRecorder(determine_wants)
624
create_pb = pb = ui.ui_factory.nested_progress_bar()
626
objects_iter = self.source.fetch_objects(
627
wants_recorder, graph_walker, store.get_raw,
629
trace.mutter("Importing %d new revisions", len(wants_recorder.wants))
630
(pack_hint, last_rev) = import_git_objects(self.target, mapping,
631
objects_iter, store, wants_recorder.wants, pb, limit)
632
return (pack_hint, last_rev, wants_recorder.remote_refs)
640
def is_compatible(source, target):
641
"""Be compatible with GitRepository."""
642
return (isinstance(source, RemoteGitRepository) and
643
target.supports_rich_root() and
644
not isinstance(target, GitRepository) and
645
target.texts is not None)
648
class InterLocalGitNonGitRepository(InterGitNonGitRepository):
649
"""InterRepository that copies revisions from a local Git into a non-Git
652
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
653
"""See `InterGitNonGitRepository`."""
654
remote_refs = self.source._git.get_refs()
655
wants = determine_wants(remote_refs)
658
create_pb = pb = ui.ui_factory.nested_progress_bar()
659
target_git_object_retriever = BazaarObjectStore(self.target, mapping)
661
self.target.lock_write()
663
(pack_hint, last_rev) = import_git_objects(self.target, mapping,
664
self.source._git.object_store,
665
target_git_object_retriever, wants, pb, limit)
666
return (pack_hint, last_rev, remote_refs)
674
def is_compatible(source, target):
675
"""Be compatible with GitRepository."""
676
return (isinstance(source, LocalGitRepository) and
677
target.supports_rich_root() and
678
not isinstance(target, GitRepository) and
679
target.texts is not None)
682
class InterGitGitRepository(InterGitRepository):
683
"""InterRepository that copies between Git repositories."""
685
def fetch_objects(self, determine_wants, mapping, pb=None):
687
trace.note("git: %s", text)
688
graphwalker = self.target._git.get_graph_walker()
689
if (isinstance(self.source, LocalGitRepository) and
690
isinstance(self.target, LocalGitRepository)):
691
refs = self.source._git.fetch(self.target._git, determine_wants,
693
return (None, None, refs)
694
elif (isinstance(self.source, LocalGitRepository) and
695
isinstance(self.target, RemoteGitRepository)):
696
raise NotImplementedError
697
elif (isinstance(self.source, RemoteGitRepository) and
698
isinstance(self.target, LocalGitRepository)):
699
f, commit = self.target._git.object_store.add_thin_pack()
701
refs = self.source.bzrdir.root_transport.fetch_pack(
702
determine_wants, graphwalker, f.write, progress)
704
return (None, None, refs)
711
def _target_has_shas(self, shas):
712
return set([sha for sha in shas if self.target._git.object_store])
714
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
715
mapping=None, fetch_spec=None, branches=None):
717
mapping = self.source.get_mapping()
719
if revision_id is not None:
720
args = [mapping.revision_id_bzr_to_foreign(revision_id)[0]]
721
elif fetch_spec is not None:
722
recipe = fetch_spec.get_recipe()
723
if recipe[0] in ("search", "proxy-search"):
726
raise AssertionError("Unsupported search result type %s" % recipe[0])
727
args = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in heads]
728
if branches is not None:
729
determine_wants = lambda x: [x[y] for y in branches if not x[y] in r.object_store and x[y] != ZERO_SHA]
730
elif fetch_spec is None and revision_id is None:
731
determine_wants = self.determine_wants_all
733
determine_wants = lambda x: [y for y in args if not y in r.object_store and y != ZERO_SHA]
734
wants_recorder = DetermineWantsRecorder(determine_wants)
735
self.fetch_objects(wants_recorder, mapping)
736
return wants_recorder.remote_refs
739
def is_compatible(source, target):
740
"""Be compatible with GitRepository."""
741
return (isinstance(source, GitRepository) and
742
isinstance(target, GitRepository))
744
def get_determine_wants_revids(self, revids, include_tags=False):
746
for revid in set(revids):
747
git_sha, mapping = self.source.lookup_bzr_revision_id(revid)
749
return self.get_determine_wants_heads(wants, include_tags=include_tags)
added
removed
fetch.py
