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)
134
and ((pkind == "symlink" and
135
ptree.get_symlink_target(ppath, file_id) ==
136
ie.symlink_target) or
138
ptree.get_file_sha1(ppath, file_id) == ie.text_sha1 and
139
ptree.is_executable(ppath, file_id) == ie.executable))):
140
# found a revision in one of the parents to use
141
ie.revision = ptree.get_file_revision(ppath, file_id)
143
parent_key = (file_id, ptree.get_file_revision(ppath, file_id))
144
if parent_key not in parent_keys:
145
parent_keys.append(parent_key)
146
if ie.revision is None:
147
# Need to store a new revision
148
ie.revision = revision_id
149
if ie.revision is None:
150
raise ValueError("no file revision set")
151
if ie.kind == 'symlink':
154
chunks = blob.chunked
155
texts.insert_record_stream([
156
ChunkedContentFactory((file_id, ie.revision),
157
tuple(parent_keys), ie.text_sha1, chunks)])
159
if base_hexsha is not None:
160
old_path = decoded_path # Renames are not supported yet
161
if stat.S_ISDIR(base_mode):
162
invdelta.extend(remove_disappeared_children(
163
base_bzr_tree, old_path, lookup_object(base_hexsha), [],
167
invdelta.append((old_path, decoded_path, file_id, ie))
168
if base_hexsha != hexsha:
169
store_updater.add_object(blob, (ie.file_id, ie.revision), path)
173
class SubmodulesRequireSubtrees(BzrError):
174
_fmt = ("The repository you are fetching from contains submodules, "
175
"which require a Bazaar format that supports tree references.")
179
def import_git_submodule(texts, mapping, path, name, hexshas,
180
base_bzr_tree, parent_id, revision_id,
181
parent_bzr_trees, lookup_object,
182
modes, store_updater, lookup_file_id):
183
"""Import a git submodule."""
184
(base_hexsha, hexsha) = hexshas
185
(base_mode, mode) = modes
186
if base_hexsha == hexsha and base_mode == mode:
188
file_id = lookup_file_id(path)
190
ie = TreeReference(file_id, name.decode("utf-8"), parent_id)
191
ie.revision = revision_id
192
if base_hexsha is not None:
193
old_path = path.decode("utf-8") # Renames are not supported yet
194
if stat.S_ISDIR(base_mode):
195
invdelta.extend(remove_disappeared_children(
196
base_bzr_tree, old_path, lookup_object(base_hexsha), [],
200
ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)
201
texts.insert_record_stream([
202
ChunkedContentFactory((file_id, ie.revision), (), None, [])])
203
invdelta.append((old_path, path, file_id, ie))
207
def remove_disappeared_children(base_bzr_tree, path, base_tree,
208
existing_children, lookup_object):
209
"""Generate an inventory delta for removed children.
211
:param base_bzr_tree: Base bzr tree against which to generate the
213
:param path: Path to process (unicode)
214
:param base_tree: Git Tree base object
215
:param existing_children: Children that still exist
216
:param lookup_object: Lookup a git object by its SHA1
217
:return: Inventory delta, as list
219
if not isinstance(path, text_type):
220
raise TypeError(path)
222
for name, mode, hexsha in base_tree.iteritems():
223
if name in existing_children:
225
c_path = posixpath.join(path, name.decode("utf-8"))
226
file_id = base_bzr_tree.path2id(c_path)
228
raise TypeError(file_id)
229
ret.append((c_path, None, file_id, None))
230
if stat.S_ISDIR(mode):
231
ret.extend(remove_disappeared_children(
232
base_bzr_tree, c_path, lookup_object(hexsha), [],
237
def import_git_tree(texts, mapping, path, name, hexshas,
238
base_bzr_tree, parent_id, revision_id, parent_bzr_trees,
239
lookup_object, modes, store_updater,
240
lookup_file_id, allow_submodules=False):
241
"""Import a git tree object into a bzr repository.
243
:param texts: VersionedFiles object to add to
244
:param path: Path in the tree (str)
245
:param name: Name of the tree (str)
246
:param tree: A git tree object
247
:param base_bzr_tree: Base inventory against which to return inventory
249
:return: Inventory delta for this subtree
251
(base_hexsha, hexsha) = hexshas
252
(base_mode, mode) = modes
253
if not isinstance(path, bytes):
254
raise TypeError(path)
255
if not isinstance(name, bytes):
256
raise TypeError(name)
257
if base_hexsha == hexsha and base_mode == mode:
258
# If nothing has changed since the base revision, we're done
261
file_id = lookup_file_id(osutils.safe_unicode(path))
262
# We just have to hope this is indeed utf-8:
263
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
264
tree = lookup_object(hexsha)
265
if base_hexsha is None:
267
old_path = None # Newly appeared here
269
base_tree = lookup_object(base_hexsha)
270
old_path = path.decode("utf-8") # Renames aren't supported yet
271
new_path = path.decode("utf-8")
272
if base_tree is None or type(base_tree) is not Tree:
273
ie.revision = revision_id
274
invdelta.append((old_path, new_path, ie.file_id, ie))
275
texts.insert_record_stream([
276
ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
277
# Remember for next time
278
existing_children = set()
280
for name, child_mode, child_hexsha in tree.iteritems():
281
existing_children.add(name)
282
child_path = posixpath.join(path, name)
283
if type(base_tree) is Tree:
285
child_base_mode, child_base_hexsha = base_tree[name]
287
child_base_hexsha = None
290
child_base_hexsha = None
292
if stat.S_ISDIR(child_mode):
293
subinvdelta, grandchildmodes = import_git_tree(
294
texts, mapping, child_path, name,
295
(child_base_hexsha, child_hexsha), base_bzr_tree, file_id,
296
revision_id, parent_bzr_trees, lookup_object,
297
(child_base_mode, child_mode), store_updater, lookup_file_id,
298
allow_submodules=allow_submodules)
299
elif S_ISGITLINK(child_mode): # submodule
300
if not allow_submodules:
301
raise SubmodulesRequireSubtrees()
302
subinvdelta, grandchildmodes = import_git_submodule(
303
texts, mapping, child_path, name,
304
(child_base_hexsha, child_hexsha),
305
base_bzr_tree, file_id, revision_id, parent_bzr_trees,
306
lookup_object, (child_base_mode, child_mode), store_updater,
309
if not mapping.is_special_file(name):
310
subinvdelta = import_git_blob(
311
texts, mapping, child_path, name,
312
(child_base_hexsha, child_hexsha), base_bzr_tree, file_id,
313
revision_id, parent_bzr_trees, lookup_object,
314
(child_base_mode, child_mode), store_updater,
319
child_modes.update(grandchildmodes)
320
invdelta.extend(subinvdelta)
321
if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
322
stat.S_IFLNK, DEFAULT_FILE_MODE | 0o111,
324
child_modes[child_path] = child_mode
325
# Remove any children that have disappeared
326
if base_tree is not None and type(base_tree) is Tree:
327
invdelta.extend(remove_disappeared_children(
328
base_bzr_tree, old_path, base_tree, existing_children,
330
store_updater.add_object(tree, (file_id, revision_id), path)
331
return invdelta, child_modes
334
def verify_commit_reconstruction(target_git_object_retriever, lookup_object,
335
o, rev, ret_tree, parent_trees, mapping,
336
unusual_modes, verifiers):
337
new_unusual_modes = mapping.export_unusual_file_modes(rev)
338
if new_unusual_modes != unusual_modes:
339
raise AssertionError("unusual modes don't match: %r != %r" % (
340
unusual_modes, new_unusual_modes))
341
# Verify that we can reconstruct the commit properly
342
rec_o = target_git_object_retriever._reconstruct_commit(rev, o.tree, True,
345
raise AssertionError("Reconstructed commit differs: %r != %r" % (
349
for path, obj, ie in _tree_to_objects(
350
ret_tree, parent_trees, target_git_object_retriever._cache.idmap,
351
unusual_modes, mapping.BZR_DUMMY_FILE):
352
old_obj_id = tree_lookup_path(lookup_object, o.tree, path)[1]
354
if obj.id != old_obj_id:
355
diff.append((path, lookup_object(old_obj_id), obj))
356
for (path, old_obj, new_obj) in diff:
357
while (old_obj.type_name == "tree"
358
and new_obj.type_name == "tree"
359
and sorted(old_obj) == sorted(new_obj)):
361
if old_obj[name][0] != new_obj[name][0]:
362
raise AssertionError(
363
"Modes for %s differ: %o != %o" %
364
(path, old_obj[name][0], new_obj[name][0]))
365
if old_obj[name][1] != new_obj[name][1]:
366
# Found a differing child, delve deeper
367
path = posixpath.join(path, name)
368
old_obj = lookup_object(old_obj[name][1])
369
new_obj = new_objs[path]
371
raise AssertionError(
372
"objects differ for %s: %r != %r" % (path, old_obj, new_obj))
375
def ensure_inventories_in_repo(repo, trees):
376
real_inv_vf = repo.inventories.without_fallbacks()
378
revid = t.get_revision_id()
379
if not real_inv_vf.get_parent_map([(revid, )]):
380
repo.add_inventory(revid, t.root_inventory, t.get_parent_ids())
383
def import_git_commit(repo, mapping, head, lookup_object,
384
target_git_object_retriever, trees_cache):
385
o = lookup_object(head)
386
# Note that this uses mapping.revision_id_foreign_to_bzr. If the parents
387
# were bzr roundtripped revisions they would be specified in the
389
rev, roundtrip_revid, verifiers = mapping.import_commit(
390
o, mapping.revision_id_foreign_to_bzr)
391
if roundtrip_revid is not None:
392
original_revid = rev.revision_id
393
rev.revision_id = roundtrip_revid
394
# We have to do this here, since we have to walk the tree and
395
# we need to make sure to import the blobs / trees with the right
396
# path; this may involve adding them more than once.
397
parent_trees = trees_cache.revision_trees(rev.parent_ids)
398
ensure_inventories_in_repo(repo, parent_trees)
399
if parent_trees == []:
400
base_bzr_tree = trees_cache.revision_tree(NULL_REVISION)
404
base_bzr_tree = parent_trees[0]
405
base_tree = lookup_object(o.parents[0]).tree
406
base_mode = stat.S_IFDIR
407
store_updater = target_git_object_retriever._get_updater(rev)
408
tree_supplement = mapping.get_fileid_map(lookup_object, o.tree)
409
inv_delta, unusual_modes = import_git_tree(
410
repo.texts, mapping, b"", b"", (base_tree, o.tree), base_bzr_tree,
411
None, rev.revision_id, parent_trees, lookup_object,
412
(base_mode, stat.S_IFDIR), store_updater,
413
tree_supplement.lookup_file_id,
414
allow_submodules=repo._format.supports_tree_reference)
415
if unusual_modes != {}:
416
for path, mode in unusual_modes.iteritems():
417
warn_unusual_mode(rev.foreign_revid, path, mode)
418
mapping.import_unusual_file_modes(rev, unusual_modes)
420
basis_id = rev.parent_ids[0]
422
basis_id = NULL_REVISION
423
base_bzr_inventory = None
425
base_bzr_inventory = base_bzr_tree.root_inventory
426
rev.inventory_sha1, inv = repo.add_inventory_by_delta(
427
basis_id, inv_delta, rev.revision_id, rev.parent_ids,
429
ret_tree = InventoryRevisionTree(repo, inv, rev.revision_id)
431
if verifiers and roundtrip_revid is not None:
432
testament = StrictTestament3(rev, ret_tree)
433
calculated_verifiers = {"testament3-sha1": testament.as_sha1()}
434
if calculated_verifiers != verifiers:
435
trace.mutter("Testament SHA1 %r for %r did not match %r.",
436
calculated_verifiers["testament3-sha1"],
437
rev.revision_id, verifiers["testament3-sha1"])
438
rev.revision_id = original_revid
439
rev.inventory_sha1, inv = repo.add_inventory_by_delta(
440
basis_id, inv_delta, rev.revision_id, rev.parent_ids,
442
ret_tree = InventoryRevisionTree(repo, inv, rev.revision_id)
444
calculated_verifiers = {}
445
store_updater.add_object(o, calculated_verifiers, None)
446
store_updater.finish()
447
trees_cache.add(ret_tree)
448
repo.add_revision(rev.revision_id, rev)
449
if "verify" in debug.debug_flags:
450
verify_commit_reconstruction(
451
target_git_object_retriever, lookup_object, o, rev, ret_tree,
452
parent_trees, mapping, unusual_modes, verifiers)
455
def import_git_objects(repo, mapping, object_iter,
456
target_git_object_retriever, heads, pb=None,
458
"""Import a set of git objects into a bzr repository.
460
:param repo: Target Bazaar repository
461
:param mapping: Mapping to use
462
:param object_iter: Iterator over Git objects.
463
:return: Tuple with pack hints and last imported revision id
465
def lookup_object(sha):
467
return object_iter[sha]
469
return target_git_object_retriever[sha]
472
heads = list(set(heads))
473
trees_cache = LRUTreeCache(repo)
474
# Find and convert commit objects
477
pb.update("finding revisions to fetch", len(graph), None)
481
if not isinstance(head, bytes):
482
raise TypeError(head)
484
o = lookup_object(head)
487
if isinstance(o, Commit):
488
rev, roundtrip_revid, verifiers = mapping.import_commit(
489
o, mapping.revision_id_foreign_to_bzr)
490
if (repo.has_revision(rev.revision_id)
491
or (roundtrip_revid and
492
repo.has_revision(roundtrip_revid))):
494
graph.append((o.id, o.parents))
495
heads.extend([p for p in o.parents if p not in checked])
496
elif isinstance(o, Tag):
497
if o.object[1] not in checked:
498
heads.append(o.object[1])
500
trace.warning("Unable to import head object %r" % o)
503
# Order the revisions
504
# Create the inventory objects
506
revision_ids = topo_sort(graph)
508
if limit is not None:
509
revision_ids = revision_ids[:limit]
511
for offset in range(0, len(revision_ids), batch_size):
512
target_git_object_retriever.start_write_group()
514
repo.start_write_group()
516
for i, head in enumerate(
517
revision_ids[offset:offset + batch_size]):
519
pb.update("fetching revisions", offset + i,
521
import_git_commit(repo, mapping, head, lookup_object,
522
target_git_object_retriever, trees_cache)
524
except BaseException:
525
repo.abort_write_group()
528
hint = repo.commit_write_group()
530
pack_hints.extend(hint)
531
except BaseException:
532
target_git_object_retriever.abort_write_group()
535
target_git_object_retriever.commit_write_group()
536
return pack_hints, last_imported
539
class DetermineWantsRecorder(object):
541
def __init__(self, actual):
544
self.remote_refs = {}
546
def __call__(self, refs):
547
if type(refs) is not dict:
548
raise TypeError(refs)
549
self.remote_refs = refs
550
self.wants = self.actual(refs)