1
# Copyright (C) 2008-2018 Jelmer Vernooij <jelmer@jelmer.uk>
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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17
"""Fetching from git into bzr."""
19
from __future__ import absolute_import
21
from dulwich.objects import (
29
from dulwich.object_store import (
41
from ..errors import (
44
from ..bzr.inventory import (
50
from ..revision import (
53
from ..bzr.inventorytree import InventoryRevisionTree
54
from ..sixish import text_type
55
from ..testament import (
61
from ..bzr.versionedfile import (
62
ChunkedContentFactory,
65
from .mapping import (
71
from .object_store import (
77
def import_git_blob(texts, mapping, path, name, hexshas,
78
base_bzr_tree, parent_id, revision_id,
79
parent_bzr_trees, lookup_object, modes, store_updater,
81
"""Import a git blob object into a bzr repository.
83
:param texts: VersionedFiles to add to
84
:param path: Path in the tree
85
:param blob: A git blob
86
:return: Inventory delta for this file
88
if not isinstance(path, bytes):
90
decoded_path = path.decode('utf-8')
91
(base_mode, mode) = modes
92
(base_hexsha, hexsha) = hexshas
93
if mapping.is_special_file(path):
95
if base_hexsha == hexsha and base_mode == mode:
96
# If nothing has changed since the base revision, we're done
98
file_id = lookup_file_id(decoded_path)
99
if stat.S_ISLNK(mode):
103
ie = cls(file_id, name.decode("utf-8"), parent_id)
104
if ie.kind == "file":
105
ie.executable = mode_is_executable(mode)
106
if base_hexsha == hexsha and mode_kind(base_mode) == mode_kind(mode):
107
base_exec = base_bzr_tree.is_executable(decoded_path)
108
if ie.kind == "symlink":
109
ie.symlink_target = base_bzr_tree.get_symlink_target(decoded_path)
111
ie.text_size = base_bzr_tree.get_file_size(decoded_path)
112
ie.text_sha1 = base_bzr_tree.get_file_sha1(decoded_path)
113
if ie.kind == "symlink" or ie.executable == base_exec:
114
ie.revision = base_bzr_tree.get_file_revision(decoded_path)
116
blob = lookup_object(hexsha)
118
blob = lookup_object(hexsha)
119
if ie.kind == "symlink":
121
ie.symlink_target = blob.data.decode("utf-8")
123
ie.text_size = sum(map(len, blob.chunked))
124
ie.text_sha1 = osutils.sha_strings(blob.chunked)
125
# Check what revision we should store
127
for ptree in parent_bzr_trees:
129
ppath = ptree.id2path(file_id)
130
except errors.NoSuchId:
132
pkind = ptree.kind(ppath, file_id)
133
if (pkind == ie.kind and
134
((pkind == "symlink" and ptree.get_symlink_target(ppath, file_id) == ie.symlink_target) or
135
(pkind == "file" and ptree.get_file_sha1(ppath, file_id) == ie.text_sha1 and
136
ptree.is_executable(ppath, file_id) == ie.executable))):
137
# found a revision in one of the parents to use
138
ie.revision = ptree.get_file_revision(ppath, file_id)
140
parent_key = (file_id, ptree.get_file_revision(ppath, file_id))
141
if not parent_key in parent_keys:
142
parent_keys.append(parent_key)
143
if ie.revision is None:
144
# Need to store a new revision
145
ie.revision = revision_id
146
if ie.revision is None:
147
raise ValueError("no file revision set")
148
if ie.kind == 'symlink':
151
chunks = blob.chunked
152
texts.insert_record_stream([
153
ChunkedContentFactory((file_id, ie.revision),
154
tuple(parent_keys), ie.text_sha1, chunks)])
156
if base_hexsha is not None:
157
old_path = decoded_path # Renames are not supported yet
158
if stat.S_ISDIR(base_mode):
159
invdelta.extend(remove_disappeared_children(base_bzr_tree, old_path,
160
lookup_object(base_hexsha), [], lookup_object))
163
invdelta.append((old_path, decoded_path, file_id, ie))
164
if base_hexsha != hexsha:
165
store_updater.add_object(blob, (ie.file_id, ie.revision), path)
169
class SubmodulesRequireSubtrees(BzrError):
170
_fmt = ("The repository you are fetching from contains submodules, "
171
"which require a Bazaar format that supports tree references.")
175
def import_git_submodule(texts, mapping, path, name, hexshas,
176
base_bzr_tree, parent_id, revision_id, parent_bzr_trees, lookup_object,
177
modes, store_updater, lookup_file_id):
178
"""Import a git submodule."""
179
(base_hexsha, hexsha) = hexshas
180
(base_mode, mode) = modes
181
if base_hexsha == hexsha and base_mode == mode:
183
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 not None:
188
old_path = path.decode("utf-8") # Renames are not supported yet
189
if stat.S_ISDIR(base_mode):
190
invdelta.extend(remove_disappeared_children(base_bzr_tree, old_path,
191
lookup_object(base_hexsha), [], lookup_object))
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.append((old_path, path, file_id, ie))
201
def remove_disappeared_children(base_bzr_tree, path, base_tree, existing_children,
203
"""Generate an inventory delta for removed children.
205
:param base_bzr_tree: Base bzr tree 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
if not isinstance(path, text_type):
214
raise TypeError(path)
216
for name, mode, hexsha in base_tree.iteritems():
217
if name in existing_children:
219
c_path = posixpath.join(path, name.decode("utf-8"))
220
file_id = base_bzr_tree.path2id(c_path)
222
raise TypeError(file_id)
223
ret.append((c_path, None, file_id, None))
224
if stat.S_ISDIR(mode):
225
ret.extend(remove_disappeared_children(
226
base_bzr_tree, c_path, lookup_object(hexsha), [], lookup_object))
230
def import_git_tree(texts, mapping, path, name, hexshas,
231
base_bzr_tree, parent_id, revision_id, parent_bzr_trees,
232
lookup_object, modes, store_updater,
233
lookup_file_id, allow_submodules=False):
234
"""Import a git tree object into a bzr repository.
236
:param texts: VersionedFiles object to add to
237
:param path: Path in the tree (str)
238
:param name: Name of the tree (str)
239
:param tree: A git tree object
240
:param base_bzr_tree: Base inventory against which to return inventory delta
241
:return: Inventory delta for this subtree
243
(base_hexsha, hexsha) = hexshas
244
(base_mode, mode) = modes
245
if not isinstance(path, bytes):
246
raise TypeError(path)
247
if not isinstance(name, bytes):
248
raise TypeError(name)
249
if base_hexsha == hexsha and base_mode == mode:
250
# If nothing has changed since the base revision, we're done
253
file_id = lookup_file_id(osutils.safe_unicode(path))
254
# We just have to hope this is indeed utf-8:
255
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
256
tree = lookup_object(hexsha)
257
if base_hexsha is None:
259
old_path = None # Newly appeared here
261
base_tree = lookup_object(base_hexsha)
262
old_path = path.decode("utf-8") # Renames aren't supported yet
263
new_path = path.decode("utf-8")
264
if base_tree is None or type(base_tree) is not Tree:
265
ie.revision = revision_id
266
invdelta.append((old_path, new_path, ie.file_id, ie))
267
texts.insert_record_stream([
268
ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
269
# Remember for next time
270
existing_children = set()
272
for name, child_mode, child_hexsha in tree.iteritems():
273
existing_children.add(name)
274
child_path = posixpath.join(path, name)
275
if type(base_tree) is Tree:
277
child_base_mode, child_base_hexsha = base_tree[name]
279
child_base_hexsha = None
282
child_base_hexsha = None
284
if stat.S_ISDIR(child_mode):
285
subinvdelta, grandchildmodes = import_git_tree(texts, mapping,
286
child_path, name, (child_base_hexsha, child_hexsha),
287
base_bzr_tree, file_id, revision_id, parent_bzr_trees,
288
lookup_object, (child_base_mode, child_mode), store_updater,
289
lookup_file_id, allow_submodules=allow_submodules)
290
elif S_ISGITLINK(child_mode): # submodule
291
if not allow_submodules:
292
raise SubmodulesRequireSubtrees()
293
subinvdelta, grandchildmodes = import_git_submodule(texts, mapping,
294
child_path, name, (child_base_hexsha, child_hexsha),
295
base_bzr_tree, file_id, revision_id, parent_bzr_trees,
296
lookup_object, (child_base_mode, child_mode), store_updater,
299
if not mapping.is_special_file(name):
300
subinvdelta = import_git_blob(texts, mapping, child_path, name,
301
(child_base_hexsha, child_hexsha), base_bzr_tree, file_id,
302
revision_id, parent_bzr_trees, lookup_object,
303
(child_base_mode, child_mode), store_updater, lookup_file_id)
307
child_modes.update(grandchildmodes)
308
invdelta.extend(subinvdelta)
309
if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
310
stat.S_IFLNK, DEFAULT_FILE_MODE|0o111,
312
child_modes[child_path] = child_mode
313
# Remove any children that have disappeared
314
if base_tree is not None and type(base_tree) is Tree:
315
invdelta.extend(remove_disappeared_children(base_bzr_tree, old_path,
316
base_tree, existing_children, lookup_object))
317
store_updater.add_object(tree, (file_id, revision_id), path)
318
return invdelta, child_modes
321
def verify_commit_reconstruction(target_git_object_retriever, lookup_object,
322
o, rev, ret_tree, parent_trees, mapping, unusual_modes, verifiers):
323
new_unusual_modes = mapping.export_unusual_file_modes(rev)
324
if new_unusual_modes != unusual_modes:
325
raise AssertionError("unusual modes don't match: %r != %r" % (
326
unusual_modes, new_unusual_modes))
327
# Verify that we can reconstruct the commit properly
328
rec_o = target_git_object_retriever._reconstruct_commit(rev, o.tree, True,
331
raise AssertionError("Reconstructed commit differs: %r != %r" % (
335
for path, obj, ie in _tree_to_objects(ret_tree, parent_trees,
336
target_git_object_retriever._cache.idmap, unusual_modes,
337
mapping.BZR_DUMMY_FILE):
338
old_obj_id = tree_lookup_path(lookup_object, o.tree, path)[1]
340
if obj.id != old_obj_id:
341
diff.append((path, lookup_object(old_obj_id), obj))
342
for (path, old_obj, new_obj) in diff:
343
while (old_obj.type_name == "tree" and
344
new_obj.type_name == "tree" and
345
sorted(old_obj) == sorted(new_obj)):
347
if old_obj[name][0] != new_obj[name][0]:
348
raise AssertionError("Modes for %s differ: %o != %o" %
349
(path, old_obj[name][0], new_obj[name][0]))
350
if old_obj[name][1] != new_obj[name][1]:
351
# Found a differing child, delve deeper
352
path = posixpath.join(path, name)
353
old_obj = lookup_object(old_obj[name][1])
354
new_obj = new_objs[path]
356
raise AssertionError("objects differ for %s: %r != %r" % (path,
360
def ensure_inventories_in_repo(repo, trees):
361
real_inv_vf = repo.inventories.without_fallbacks()
363
revid = t.get_revision_id()
364
if not real_inv_vf.get_parent_map([(revid, )]):
365
repo.add_inventory(revid, t.root_inventory, t.get_parent_ids())
368
def import_git_commit(repo, mapping, head, lookup_object,
369
target_git_object_retriever, trees_cache):
370
o = lookup_object(head)
371
# Note that this uses mapping.revision_id_foreign_to_bzr. If the parents
372
# were bzr roundtripped revisions they would be specified in the
374
rev, roundtrip_revid, verifiers = mapping.import_commit(
375
o, mapping.revision_id_foreign_to_bzr)
376
if roundtrip_revid is not None:
377
original_revid = rev.revision_id
378
rev.revision_id = roundtrip_revid
379
# We have to do this here, since we have to walk the tree and
380
# we need to make sure to import the blobs / trees with the right
381
# path; this may involve adding them more than once.
382
parent_trees = trees_cache.revision_trees(rev.parent_ids)
383
ensure_inventories_in_repo(repo, parent_trees)
384
if parent_trees == []:
385
base_bzr_tree = trees_cache.revision_tree(NULL_REVISION)
389
base_bzr_tree = parent_trees[0]
390
base_tree = lookup_object(o.parents[0]).tree
391
base_mode = stat.S_IFDIR
392
store_updater = target_git_object_retriever._get_updater(rev)
393
tree_supplement = mapping.get_fileid_map(lookup_object, o.tree)
394
inv_delta, unusual_modes = import_git_tree(repo.texts,
395
mapping, b"", b"", (base_tree, o.tree), base_bzr_tree,
396
None, rev.revision_id, parent_trees,
397
lookup_object, (base_mode, stat.S_IFDIR), store_updater,
398
tree_supplement.lookup_file_id,
399
allow_submodules=repo._format.supports_tree_reference)
400
if unusual_modes != {}:
401
for path, mode in unusual_modes.iteritems():
402
warn_unusual_mode(rev.foreign_revid, path, mode)
403
mapping.import_unusual_file_modes(rev, unusual_modes)
405
basis_id = rev.parent_ids[0]
407
basis_id = NULL_REVISION
408
base_bzr_inventory = None
410
base_bzr_inventory = base_bzr_tree.root_inventory
411
rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,
412
inv_delta, rev.revision_id, rev.parent_ids,
414
ret_tree = InventoryRevisionTree(repo, inv, rev.revision_id)
416
if verifiers and roundtrip_revid is not None:
417
testament = StrictTestament3(rev, ret_tree)
418
calculated_verifiers = { "testament3-sha1": testament.as_sha1() }
419
if calculated_verifiers != verifiers:
420
trace.mutter("Testament SHA1 %r for %r did not match %r.",
421
calculated_verifiers["testament3-sha1"],
422
rev.revision_id, verifiers["testament3-sha1"])
423
rev.revision_id = original_revid
424
rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,
425
inv_delta, rev.revision_id, rev.parent_ids, base_bzr_tree)
426
ret_tree = InventoryRevisionTree(repo, inv, rev.revision_id)
428
calculated_verifiers = {}
429
store_updater.add_object(o, calculated_verifiers, None)
430
store_updater.finish()
431
trees_cache.add(ret_tree)
432
repo.add_revision(rev.revision_id, rev)
433
if "verify" in debug.debug_flags:
434
verify_commit_reconstruction(target_git_object_retriever,
435
lookup_object, o, rev, ret_tree, parent_trees, mapping,
436
unusual_modes, verifiers)
439
def import_git_objects(repo, mapping, object_iter,
440
target_git_object_retriever, heads, pb=None, limit=None):
441
"""Import a set of git objects into a bzr repository.
443
:param repo: Target Bazaar repository
444
:param mapping: Mapping to use
445
:param object_iter: Iterator over Git objects.
446
:return: Tuple with pack hints and last imported revision id
448
def lookup_object(sha):
450
return object_iter[sha]
452
return target_git_object_retriever[sha]
455
heads = list(set(heads))
456
trees_cache = LRUTreeCache(repo)
457
# Find and convert commit objects
460
pb.update("finding revisions to fetch", len(graph), None)
464
if not isinstance(head, bytes):
465
raise TypeError(head)
467
o = lookup_object(head)
470
if isinstance(o, Commit):
471
rev, roundtrip_revid, verifiers = mapping.import_commit(o,
472
mapping.revision_id_foreign_to_bzr)
473
if (repo.has_revision(rev.revision_id) or
474
(roundtrip_revid and repo.has_revision(roundtrip_revid))):
476
graph.append((o.id, o.parents))
477
heads.extend([p for p in o.parents if p not in checked])
478
elif isinstance(o, Tag):
479
if o.object[1] not in checked:
480
heads.append(o.object[1])
482
trace.warning("Unable to import head object %r" % o)
485
# Order the revisions
486
# Create the inventory objects
488
revision_ids = topo_sort(graph)
490
if limit is not None:
491
revision_ids = revision_ids[:limit]
493
for offset in range(0, len(revision_ids), batch_size):
494
target_git_object_retriever.start_write_group()
496
repo.start_write_group()
498
for i, head in enumerate(
499
revision_ids[offset:offset+batch_size]):
501
pb.update("fetching revisions", offset+i,
503
import_git_commit(repo, mapping, head, lookup_object,
504
target_git_object_retriever, trees_cache)
507
repo.abort_write_group()
510
hint = repo.commit_write_group()
512
pack_hints.extend(hint)
514
target_git_object_retriever.abort_write_group()
517
target_git_object_retriever.commit_write_group()
518
return pack_hints, last_imported
521
class DetermineWantsRecorder(object):
523
def __init__(self, actual):
526
self.remote_refs = {}
528
def __call__(self, refs):
529
if type(refs) is not dict:
530
raise TypeError(refs)
531
self.remote_refs = refs
532
self.wants = self.actual(refs)