1
# Copyright (C) 2008 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 (
36
from bzrlib.errors import (
40
from bzrlib.inventory import (
47
from bzrlib.repository import (
50
from bzrlib.revision import (
53
from bzrlib.tsort import (
56
from bzrlib.versionedfile import (
57
ChunkedContentFactory,
60
from bzrlib.plugins.git.mapping import (
62
inventory_to_tree_and_blobs,
68
from bzrlib.plugins.git.object_store import (
72
from bzrlib.plugins.git.remote import (
75
from bzrlib.plugins.git.repository import (
82
def import_git_blob(texts, mapping, path, name, (base_hexsha, hexsha),
83
base_inv, base_inv_shamap, parent_id, revision_id,
84
parent_invs, lookup_object, (base_mode, mode)):
85
"""Import a git blob object into a bzr repository.
87
:param texts: VersionedFiles to add to
88
:param path: Path in the tree
89
:param blob: A git blob
90
:return: Inventory delta for this file
92
if base_hexsha == hexsha and base_mode == mode:
93
# If nothing has changed since the base revision, we're done
95
file_id = mapping.generate_file_id(path)
96
if stat.S_ISLNK(mode):
100
ie = cls(file_id, name.decode("utf-8"), parent_id)
101
ie.executable = mode_is_executable(mode)
102
if base_hexsha == hexsha and mode_kind(base_mode) == mode_kind(mode):
103
base_ie = base_inv[base_inv.path2id(path)]
104
ie.text_size = base_ie.text_size
105
ie.text_sha1 = base_ie.text_sha1
106
ie.symlink_target = base_ie.symlink_target
107
if ie.executable == base_ie.executable:
108
ie.revision = base_ie.revision
110
blob = lookup_object(hexsha)
112
blob = lookup_object(hexsha)
113
if ie.kind == "symlink":
115
ie.symlink_target = blob.data
119
ie.text_size = len(blob.data)
120
ie.text_sha1 = osutils.sha_string(blob.data)
121
# Check what revision we should store
123
for pinv in parent_invs[1:]:
124
if pinv.revision_id == base_inv.revision_id:
133
if pie.text_sha1 == ie.text_sha1 and pie.executable == ie.executable and pie.symlink_target == ie.symlink_target:
134
# found a revision in one of the parents to use
135
ie.revision = pie.revision
137
parent_keys.append((file_id, pie.revision))
138
if ie.revision is None:
139
# Need to store a new revision
140
ie.revision = revision_id
141
assert ie.revision is not None
142
if ie.kind == 'symlink':
146
chunks = blob.chunked
147
except AttributeError: # older version of dulwich
149
texts.insert_record_stream([ChunkedContentFactory((file_id, ie.revision), tuple(parent_keys), ie.text_sha1, chunks)])
151
if base_hexsha is not None:
152
old_path = path # Renames are not supported yet
153
if stat.S_ISDIR(base_mode):
154
invdelta.extend(remove_disappeared_children(base_inv, old_path,
155
lookup_object(base_hexsha), [], lookup_object))
158
invdelta.append((old_path, path, file_id, ie))
159
return (invdelta, [(ie.file_id, "blob", hexsha, ie.revision)])
162
class SubmodulesRequireSubtrees(BzrError):
163
_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'."""
167
def import_git_submodule(texts, mapping, path, name, (base_hexsha, hexsha),
168
base_inv, parent_id, revision_id, parent_invs, lookup_object,
170
if base_hexsha == hexsha and base_mode == mode:
172
file_id = mapping.generate_file_id(path)
173
ie = TreeReference(file_id, name.decode("utf-8"), parent_id)
174
ie.revision = revision_id
175
if base_hexsha is None:
179
ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)
180
texts.insert_record_stream([ChunkedContentFactory((file_id, ie.revision), (), None, [])])
181
invdelta = [(oldpath, path, file_id, ie)]
182
return invdelta, {}, []
185
def remove_disappeared_children(base_inv, path, base_tree, existing_children,
188
for name, mode, hexsha in base_tree.iteritems():
189
if name in existing_children:
191
c_path = posixpath.join(path, name.decode("utf-8"))
192
ret.append((c_path, None, base_inv.path2id(c_path), None))
193
if stat.S_ISDIR(mode):
194
ret.extend(remove_disappeared_children(
195
base_inv, c_path, lookup_object(hexsha), [], lookup_object))
199
def import_git_tree(texts, mapping, path, name, (base_hexsha, hexsha),
200
base_inv, base_inv_shamap, parent_id, revision_id, parent_invs,
201
lookup_object, (base_mode, mode), allow_submodules=False):
202
"""Import a git tree object into a bzr repository.
204
:param texts: VersionedFiles object to add to
205
:param path: Path in the tree
206
:param tree: A git tree object
207
:param base_inv: Base inventory against which to return inventory delta
208
:return: Inventory delta for this subtree
210
if base_hexsha == hexsha and base_mode == mode:
211
# If nothing has changed since the base revision, we're done
214
file_id = mapping.generate_file_id(path)
215
# We just have to hope this is indeed utf-8:
216
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
217
tree = lookup_object(hexsha)
218
if base_hexsha is None:
220
old_path = None # Newly appeared here
222
base_tree = lookup_object(base_hexsha)
223
old_path = path # Renames aren't supported yet
224
if base_tree is None or type(base_tree) is not Tree:
225
ie.revision = revision_id
226
invdelta.append((old_path, path, ie.file_id, ie))
227
texts.insert_record_stream([ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
228
# Remember for next time
229
existing_children = set()
231
shamap = [(ie.file_id, "tree", hexsha, revision_id)]
232
for child_mode, name, child_hexsha in tree.entries():
233
existing_children.add(name)
234
child_path = posixpath.join(path, name)
235
if type(base_tree) is Tree:
237
child_base_mode, child_base_hexsha = base_tree[name]
239
child_base_hexsha = None
242
child_base_hexsha = None
244
if stat.S_ISDIR(child_mode):
245
subinvdelta, grandchildmodes, subshamap = import_git_tree(
246
texts, mapping, child_path, name,
247
(child_base_hexsha, child_hexsha),
248
base_inv, base_inv_shamap,
249
file_id, revision_id, parent_invs, lookup_object,
250
(child_base_mode, child_mode),
251
allow_submodules=allow_submodules)
252
elif S_ISGITLINK(child_mode): # submodule
253
if not allow_submodules:
254
raise SubmodulesRequireSubtrees()
255
subinvdelta, grandchildmodes, subshamap = import_git_submodule(
256
texts, mapping, child_path, name,
257
(child_base_hexsha, child_hexsha),
258
base_inv, file_id, revision_id, parent_invs, lookup_object,
259
(child_base_mode, child_mode))
261
subinvdelta, subshamap = import_git_blob(texts, mapping,
262
child_path, name, (child_base_hexsha, child_hexsha),
263
base_inv, base_inv_shamap,
265
revision_id, parent_invs, lookup_object,
266
(child_base_mode, child_mode))
268
child_modes.update(grandchildmodes)
269
invdelta.extend(subinvdelta)
270
shamap.extend(subshamap)
271
if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
272
stat.S_IFLNK, DEFAULT_FILE_MODE|0111):
273
child_modes[child_path] = child_mode
274
# Remove any children that have disappeared
275
if base_tree is not None and type(base_tree) is Tree:
276
invdelta.extend(remove_disappeared_children(base_inv, old_path,
277
base_tree, existing_children, lookup_object))
278
return invdelta, child_modes, shamap
281
def import_git_commit(repo, mapping, head, lookup_object,
282
target_git_object_retriever, parent_invs_cache):
283
o = lookup_object(head)
284
rev = mapping.import_commit(o)
285
# We have to do this here, since we have to walk the tree and
286
# we need to make sure to import the blobs / trees with the right
287
# path; this may involve adding them more than once.
288
parent_invs = parent_invs_cache.get_inventories(rev.parent_ids)
289
if parent_invs == []:
290
base_inv = Inventory(root_id=None)
291
base_inv_shamap = None # Should never be accessed
295
base_inv = parent_invs[0]
296
base_inv_shamap = target_git_object_retriever._idmap.get_inventory_sha_map(base_inv.revision_id)
297
base_tree = lookup_object(o.parents[0]).tree
298
base_mode = stat.S_IFDIR
299
inv_delta, unusual_modes, entries = import_git_tree(repo.texts,
300
mapping, "", u"", (base_tree, o.tree), base_inv, base_inv_shamap,
301
None, rev.revision_id, parent_invs, lookup_object,
302
(base_mode, stat.S_IFDIR),
303
allow_submodules=getattr(repo._format, "supports_tree_reference", False))
304
target_git_object_retriever._idmap.add_entries(rev.revision_id,
305
rev.parent_ids, head, o.tree, entries)
306
if unusual_modes != {}:
307
for path, mode in unusual_modes.iteritems():
308
warn_unusual_mode(rev.foreign_revid, path, mode)
309
mapping.import_unusual_file_modes(rev, unusual_modes)
311
basis_id = rev.parent_ids[0]
313
basis_id = NULL_REVISION
315
rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,
316
inv_delta, rev.revision_id, rev.parent_ids,
318
parent_invs_cache.add(rev.revision_id, inv)
319
repo.add_revision(rev.revision_id, rev)
320
if "verify" in debug.debug_flags:
321
new_unusual_modes = mapping.export_unusual_file_modes(rev)
322
if new_unusual_modes != unusual_modes:
323
raise AssertionError("unusual modes don't match: %r != %r" % (unusual_modes, new_unusual_modes))
324
objs = inventory_to_tree_and_blobs(inv, repo.texts, mapping, unusual_modes)
325
for newsha1, newobj, path in objs:
326
assert path is not None
330
(oldmode, oldsha1) = tree_lookup_path(lookup_object, o.tree, path)
331
if oldsha1 != newsha1:
332
raise AssertionError("%r != %r in %s" % (oldsha1, newsha1, path))
335
def import_git_objects(repo, mapping, object_iter,
336
target_git_object_retriever, heads, pb=None, limit=None):
337
"""Import a set of git objects into a bzr repository.
339
:param repo: Target Bazaar repository
340
:param mapping: Mapping to use
341
:param object_iter: Iterator over Git objects.
342
:return: Tuple with pack hints and last imported revision id
344
def lookup_object(sha):
346
return object_iter[sha]
348
return target_git_object_retriever[sha]
351
heads = list(set(heads))
352
parent_invs_cache = LRUInventoryCache(repo)
353
# Find and convert commit objects
356
pb.update("finding revisions to fetch", len(graph), None)
358
assert isinstance(head, str)
360
o = lookup_object(head)
363
if isinstance(o, Commit):
364
rev = mapping.import_commit(o)
365
if repo.has_revision(rev.revision_id):
367
squash_revision(repo, rev)
368
graph.append((o.id, o.parents))
369
heads.extend([p for p in o.parents if p not in checked])
370
elif isinstance(o, Tag):
371
if o.object[1] not in checked:
372
heads.append(o.object[1])
374
trace.warning("Unable to import head object %r" % o)
377
# Order the revisions
378
# Create the inventory objects
380
revision_ids = topo_sort(graph)
382
if limit is not None:
383
revision_ids = revision_ids[:limit]
385
for offset in range(0, len(revision_ids), batch_size):
386
target_git_object_retriever.start_write_group() # FIXME: try/finally
388
repo.start_write_group()
390
for i, head in enumerate(
391
revision_ids[offset:offset+batch_size]):
393
pb.update("fetching revisions", offset+i,
395
import_git_commit(repo, mapping, head, lookup_object,
396
target_git_object_retriever,
400
repo.abort_write_group()
403
hint = repo.commit_write_group()
405
pack_hints.extend(hint)
407
target_git_object_retriever.abort_write_group()
410
target_git_object_retriever.commit_write_group()
411
return pack_hints, last_imported
414
class InterGitRepository(InterRepository):
416
_matching_repo_format = GitRepositoryFormat()
419
def _get_repo_format_to_test():
422
def copy_content(self, revision_id=None, pb=None):
423
"""See InterRepository.copy_content."""
424
self.fetch(revision_id, pb, find_ghosts=False)
426
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
427
mapping=None, fetch_spec=None):
428
self.fetch_refs(revision_id=revision_id, pb=pb,
429
find_ghosts=find_ghosts, mapping=mapping, fetch_spec=fetch_spec)
432
class InterGitNonGitRepository(InterGitRepository):
433
"""Base InterRepository that copies revisions from a Git into a non-Git
436
def fetch_refs(self, revision_id=None, pb=None, find_ghosts=False,
437
mapping=None, fetch_spec=None):
439
mapping = self.source.get_mapping()
440
if revision_id is not None:
441
interesting_heads = [revision_id]
442
elif fetch_spec is not None:
443
interesting_heads = fetch_spec.heads
445
interesting_heads = None
447
def determine_wants(refs):
449
if interesting_heads is None:
450
ret = [sha for (ref, sha) in refs.iteritems() if not ref.endswith("^{}")]
452
ret = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in interesting_heads if revid not in (None, NULL_REVISION)]
453
return [rev for rev in ret if not self.target.has_revision(mapping.revision_id_foreign_to_bzr(rev))]
454
(pack_hint, _) = self.fetch_objects(determine_wants, mapping, pb)
455
if pack_hint is not None and self.target._format.pack_compresses:
456
self.target.pack(hint=pack_hint)
457
if interesting_heads is not None:
458
present_interesting_heads = self.target.has_revisions(interesting_heads)
459
missing_interesting_heads = set(interesting_heads) - present_interesting_heads
460
if missing_interesting_heads:
461
raise AssertionError("Missing interesting heads: %r" % missing_interesting_heads)
465
_GIT_PROGRESS_RE = re.compile(r"(.*?): +(\d+)% \((\d+)/(\d+)\)")
466
def report_git_progress(pb, text):
467
text = text.rstrip("\r\n")
468
g = _GIT_PROGRESS_RE.match(text)
470
(text, pct, current, total) = g.groups()
471
pb.update(text, int(current), int(total))
473
pb.update(text, 0, 0)
476
class InterRemoteGitNonGitRepository(InterGitNonGitRepository):
477
"""InterRepository that copies revisions from a remote Git into a non-Git
480
def get_target_heads(self):
481
# FIXME: This should be more efficient
482
all_revs = self.target.all_revision_ids()
483
parent_map = self.target.get_parent_map(all_revs)
485
map(all_parents.update, parent_map.itervalues())
486
return set(all_revs) - all_parents
488
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
490
report_git_progress(pb, text)
491
store = BazaarObjectStore(self.target, mapping)
492
self.target.lock_write()
494
heads = self.get_target_heads()
495
graph_walker = store.get_graph_walker(
496
[store._lookup_revision_sha1(head) for head in heads])
499
def record_determine_wants(heads):
500
wants = determine_wants(heads)
501
recorded_wants.extend(wants)
506
create_pb = pb = ui.ui_factory.nested_progress_bar()
508
objects_iter = self.source.fetch_objects(
509
record_determine_wants, graph_walker,
510
store.get_raw, progress)
511
return import_git_objects(self.target, mapping,
512
objects_iter, store, recorded_wants, pb, limit)
520
def is_compatible(source, target):
521
"""Be compatible with GitRepository."""
522
# FIXME: Also check target uses VersionedFile
523
return (isinstance(source, RemoteGitRepository) and
524
target.supports_rich_root() and
525
not isinstance(target, GitRepository))
528
class InterLocalGitNonGitRepository(InterGitNonGitRepository):
529
"""InterRepository that copies revisions from a local Git into a non-Git
532
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
535
wants = determine_wants(self.source._git.get_refs())
538
create_pb = pb = ui.ui_factory.nested_progress_bar()
539
target_git_object_retriever = BazaarObjectStore(self.target, mapping)
541
self.target.lock_write()
543
return import_git_objects(self.target, mapping,
544
self.source._git.object_store,
545
target_git_object_retriever, wants, pb, limit)
553
def is_compatible(source, target):
554
"""Be compatible with GitRepository."""
555
# FIXME: Also check target uses VersionedFile
556
return (isinstance(source, LocalGitRepository) and
557
target.supports_rich_root() and
558
not isinstance(target, GitRepository))
561
class InterGitGitRepository(InterGitRepository):
562
"""InterRepository that copies between Git repositories."""
564
def fetch_objects(self, determine_wants, mapping, pb=None):
566
trace.note("git: %s", text)
567
graphwalker = self.target._git.get_graph_walker()
568
if (isinstance(self.source, LocalGitRepository) and
569
isinstance(self.target, LocalGitRepository)):
570
return self.source._git.fetch(self.target._git, determine_wants,
572
elif (isinstance(self.source, LocalGitRepository) and
573
isinstance(self.target, RemoteGitRepository)):
574
raise NotImplementedError
575
elif (isinstance(self.source, RemoteGitRepository) and
576
isinstance(self.target, LocalGitRepository)):
577
f, commit = self.target._git.object_store.add_thin_pack()
579
refs = self.source._git.fetch_pack(determine_wants,
580
graphwalker, f.write, progress)
589
def fetch_refs(self, revision_id=None, pb=None, find_ghosts=False,
590
mapping=None, fetch_spec=None, branches=None):
592
mapping = self.source.get_mapping()
594
if revision_id is not None:
595
args = [mapping.revision_id_bzr_to_foreign(revision_id)[0]]
596
elif fetch_spec is not None:
597
args = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in fetch_spec.heads]
598
if branches is not None:
599
determine_wants = lambda x: [x[y] for y in branches if not x[y] in r.object_store]
600
elif fetch_spec is None and revision_id is None:
601
determine_wants = r.object_store.determine_wants_all
603
determine_wants = lambda x: [y for y in args if not y in r.object_store]
604
return self.fetch_objects(determine_wants, mapping)[0]
608
def is_compatible(source, target):
609
"""Be compatible with GitRepository."""
610
return (isinstance(source, GitRepository) and
611
isinstance(target, GitRepository))
added
removed
fetch.py
