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 ..bzr.testament import (
58
from ..tree import InterTree
62
from ..bzr.versionedfile import (
63
ChunkedContentFactory,
66
from .mapping import (
72
from .object_store import (
78
def import_git_blob(texts, mapping, path, name, hexshas,
79
base_bzr_tree, parent_id, revision_id,
80
parent_bzr_trees, lookup_object, modes, store_updater,
82
"""Import a git blob object into a bzr repository.
84
:param texts: VersionedFiles to add to
85
:param path: Path in the tree
86
:param blob: A git blob
87
:return: Inventory delta for this file
89
if not isinstance(path, bytes):
91
decoded_path = path.decode('utf-8')
92
(base_mode, mode) = modes
93
(base_hexsha, hexsha) = hexshas
94
if mapping.is_special_file(path):
96
if base_hexsha == hexsha and base_mode == mode:
97
# If nothing has changed since the base revision, we're done
99
file_id = lookup_file_id(decoded_path)
100
if stat.S_ISLNK(mode):
104
ie = cls(file_id, name.decode("utf-8"), parent_id)
105
if ie.kind == "file":
106
ie.executable = mode_is_executable(mode)
107
if base_hexsha == hexsha and mode_kind(base_mode) == mode_kind(mode):
108
base_exec = base_bzr_tree.is_executable(decoded_path)
109
if ie.kind == "symlink":
110
ie.symlink_target = base_bzr_tree.get_symlink_target(decoded_path)
112
ie.text_size = base_bzr_tree.get_file_size(decoded_path)
113
ie.text_sha1 = base_bzr_tree.get_file_sha1(decoded_path)
114
if ie.kind == "symlink" or ie.executable == base_exec:
115
ie.revision = base_bzr_tree.get_file_revision(decoded_path)
117
blob = lookup_object(hexsha)
119
blob = lookup_object(hexsha)
120
if ie.kind == "symlink":
122
ie.symlink_target = blob.data.decode("utf-8")
124
ie.text_size = sum(map(len, blob.chunked))
125
ie.text_sha1 = osutils.sha_strings(blob.chunked)
126
# Check what revision we should store
128
for ptree in parent_bzr_trees:
129
intertree = InterTree.get(ptree, base_bzr_tree)
131
ppath = intertree.find_source_paths(decoded_path, recurse='none')
132
except errors.NoSuchFile:
136
pkind = ptree.kind(ppath)
137
if (pkind == ie.kind and
138
((pkind == "symlink" and ptree.get_symlink_target(ppath) == ie.symlink_target) or
139
(pkind == "file" and ptree.get_file_sha1(ppath) == ie.text_sha1 and
140
ptree.is_executable(ppath) == ie.executable))):
141
# found a revision in one of the parents to use
142
ie.revision = ptree.get_file_revision(ppath)
144
parent_key = (file_id, ptree.get_file_revision(ppath))
145
if parent_key not in parent_keys:
146
parent_keys.append(parent_key)
147
if ie.revision is None:
148
# Need to store a new revision
149
ie.revision = revision_id
150
if ie.revision is None:
151
raise ValueError("no file revision set")
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 = decoded_path # Renames are not supported yet
162
if stat.S_ISDIR(base_mode):
163
invdelta.extend(remove_disappeared_children(
164
base_bzr_tree, old_path, lookup_object(base_hexsha), [],
168
invdelta.append((old_path, decoded_path, file_id, ie))
169
if base_hexsha != hexsha:
170
store_updater.add_object(blob, (ie.file_id, ie.revision), path)
174
class SubmodulesRequireSubtrees(BzrError):
175
_fmt = ("The repository you are fetching from contains submodules, "
176
"which require a Bazaar format that supports tree references.")
180
def import_git_submodule(texts, mapping, path, name, hexshas,
181
base_bzr_tree, parent_id, revision_id,
182
parent_bzr_trees, lookup_object,
183
modes, store_updater, lookup_file_id):
184
"""Import a git submodule."""
185
(base_hexsha, hexsha) = hexshas
186
(base_mode, mode) = modes
187
if base_hexsha == hexsha and base_mode == mode:
189
path = path.decode('utf-8')
190
file_id = lookup_file_id(path)
192
ie = TreeReference(file_id, name.decode("utf-8"), parent_id)
193
ie.revision = revision_id
194
if base_hexsha is not None:
195
old_path = path # Renames are not supported yet
196
if stat.S_ISDIR(base_mode):
197
invdelta.extend(remove_disappeared_children(
198
base_bzr_tree, old_path, lookup_object(base_hexsha), [],
202
ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)
203
texts.insert_record_stream([
204
ChunkedContentFactory((file_id, ie.revision), (), None, [])])
205
invdelta.append((old_path, path, file_id, ie))
209
def remove_disappeared_children(base_bzr_tree, path, base_tree,
210
existing_children, lookup_object):
211
"""Generate an inventory delta for removed children.
213
:param base_bzr_tree: Base bzr tree against which to generate the
215
:param path: Path to process (unicode)
216
:param base_tree: Git Tree base object
217
:param existing_children: Children that still exist
218
:param lookup_object: Lookup a git object by its SHA1
219
:return: Inventory delta, as list
221
if not isinstance(path, text_type):
222
raise TypeError(path)
224
for name, mode, hexsha in base_tree.iteritems():
225
if name in existing_children:
227
c_path = posixpath.join(path, name.decode("utf-8"))
228
file_id = base_bzr_tree.path2id(c_path)
230
raise TypeError(file_id)
231
ret.append((c_path, None, file_id, None))
232
if stat.S_ISDIR(mode):
233
ret.extend(remove_disappeared_children(
234
base_bzr_tree, c_path, lookup_object(hexsha), [],
239
def import_git_tree(texts, mapping, path, name, hexshas,
240
base_bzr_tree, parent_id, revision_id, parent_bzr_trees,
241
lookup_object, modes, store_updater,
242
lookup_file_id, allow_submodules=False):
243
"""Import a git tree object into a bzr repository.
245
:param texts: VersionedFiles object to add to
246
:param path: Path in the tree (str)
247
:param name: Name of the tree (str)
248
:param tree: A git tree object
249
:param base_bzr_tree: Base inventory against which to return inventory
251
:return: Inventory delta for this subtree
253
(base_hexsha, hexsha) = hexshas
254
(base_mode, mode) = modes
255
if not isinstance(path, bytes):
256
raise TypeError(path)
257
if not isinstance(name, bytes):
258
raise TypeError(name)
259
if base_hexsha == hexsha and base_mode == mode:
260
# If nothing has changed since the base revision, we're done
263
file_id = lookup_file_id(osutils.safe_unicode(path))
264
# We just have to hope this is indeed utf-8:
265
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
266
tree = lookup_object(hexsha)
267
if base_hexsha is None:
269
old_path = None # Newly appeared here
271
base_tree = lookup_object(base_hexsha)
272
old_path = path.decode("utf-8") # Renames aren't supported yet
273
new_path = path.decode("utf-8")
274
if base_tree is None or type(base_tree) is not Tree:
275
ie.revision = revision_id
276
invdelta.append((old_path, new_path, ie.file_id, ie))
277
texts.insert_record_stream([
278
ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
279
# Remember for next time
280
existing_children = set()
282
for name, child_mode, child_hexsha in tree.iteritems():
283
existing_children.add(name)
284
child_path = posixpath.join(path, name)
285
if type(base_tree) is Tree:
287
child_base_mode, child_base_hexsha = base_tree[name]
289
child_base_hexsha = None
292
child_base_hexsha = None
294
if stat.S_ISDIR(child_mode):
295
subinvdelta, grandchildmodes = import_git_tree(
296
texts, mapping, child_path, name,
297
(child_base_hexsha, child_hexsha), base_bzr_tree, file_id,
298
revision_id, parent_bzr_trees, lookup_object,
299
(child_base_mode, child_mode), store_updater, lookup_file_id,
300
allow_submodules=allow_submodules)
301
elif S_ISGITLINK(child_mode): # submodule
302
if not allow_submodules:
303
raise SubmodulesRequireSubtrees()
304
subinvdelta, grandchildmodes = import_git_submodule(
305
texts, mapping, child_path, name,
306
(child_base_hexsha, child_hexsha),
307
base_bzr_tree, file_id, revision_id, parent_bzr_trees,
308
lookup_object, (child_base_mode, child_mode), store_updater,
311
if not mapping.is_special_file(name):
312
subinvdelta = import_git_blob(
313
texts, mapping, child_path, name,
314
(child_base_hexsha, child_hexsha), base_bzr_tree, file_id,
315
revision_id, parent_bzr_trees, lookup_object,
316
(child_base_mode, child_mode), store_updater,
321
child_modes.update(grandchildmodes)
322
invdelta.extend(subinvdelta)
323
if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
324
stat.S_IFLNK, DEFAULT_FILE_MODE | 0o111,
326
child_modes[child_path] = child_mode
327
# Remove any children that have disappeared
328
if base_tree is not None and type(base_tree) is Tree:
329
invdelta.extend(remove_disappeared_children(
330
base_bzr_tree, old_path, base_tree, existing_children,
332
store_updater.add_object(tree, (file_id, revision_id), path)
333
return invdelta, child_modes
336
def verify_commit_reconstruction(target_git_object_retriever, lookup_object,
337
o, rev, ret_tree, parent_trees, mapping,
338
unusual_modes, verifiers):
339
new_unusual_modes = mapping.export_unusual_file_modes(rev)
340
if new_unusual_modes != unusual_modes:
341
raise AssertionError("unusual modes don't match: %r != %r" % (
342
unusual_modes, new_unusual_modes))
343
# Verify that we can reconstruct the commit properly
344
rec_o = target_git_object_retriever._reconstruct_commit(rev, o.tree, True,
347
raise AssertionError("Reconstructed commit differs: %r != %r" % (
351
for path, obj, ie in _tree_to_objects(
352
ret_tree, parent_trees, target_git_object_retriever._cache.idmap,
353
unusual_modes, mapping.BZR_DUMMY_FILE):
354
old_obj_id = tree_lookup_path(lookup_object, o.tree, path)[1]
356
if obj.id != old_obj_id:
357
diff.append((path, lookup_object(old_obj_id), obj))
358
for (path, old_obj, new_obj) in diff:
359
while (old_obj.type_name == "tree"
360
and new_obj.type_name == "tree"
361
and sorted(old_obj) == sorted(new_obj)):
363
if old_obj[name][0] != new_obj[name][0]:
364
raise AssertionError(
365
"Modes for %s differ: %o != %o" %
366
(path, old_obj[name][0], new_obj[name][0]))
367
if old_obj[name][1] != new_obj[name][1]:
368
# Found a differing child, delve deeper
369
path = posixpath.join(path, name)
370
old_obj = lookup_object(old_obj[name][1])
371
new_obj = new_objs[path]
373
raise AssertionError(
374
"objects differ for %s: %r != %r" % (path, old_obj, new_obj))
377
def ensure_inventories_in_repo(repo, trees):
378
real_inv_vf = repo.inventories.without_fallbacks()
380
revid = t.get_revision_id()
381
if not real_inv_vf.get_parent_map([(revid, )]):
382
repo.add_inventory(revid, t.root_inventory, t.get_parent_ids())
385
def import_git_commit(repo, mapping, head, lookup_object,
386
target_git_object_retriever, trees_cache, strict):
387
o = lookup_object(head)
388
# Note that this uses mapping.revision_id_foreign_to_bzr. If the parents
389
# were bzr roundtripped revisions they would be specified in the
391
rev, roundtrip_revid, verifiers = mapping.import_commit(
392
o, mapping.revision_id_foreign_to_bzr, strict)
393
if roundtrip_revid is not None:
394
original_revid = rev.revision_id
395
rev.revision_id = roundtrip_revid
396
# We have to do this here, since we have to walk the tree and
397
# we need to make sure to import the blobs / trees with the right
398
# path; this may involve adding them more than once.
399
parent_trees = trees_cache.revision_trees(rev.parent_ids)
400
ensure_inventories_in_repo(repo, parent_trees)
401
if parent_trees == []:
402
base_bzr_tree = trees_cache.revision_tree(NULL_REVISION)
406
base_bzr_tree = parent_trees[0]
407
base_tree = lookup_object(o.parents[0]).tree
408
base_mode = stat.S_IFDIR
409
store_updater = target_git_object_retriever._get_updater(rev)
410
inv_delta, unusual_modes = import_git_tree(
411
repo.texts, mapping, b"", b"", (base_tree, o.tree), base_bzr_tree,
412
None, rev.revision_id, parent_trees, lookup_object,
413
(base_mode, stat.S_IFDIR), store_updater,
414
mapping.generate_file_id,
415
allow_submodules=repo._format.supports_tree_reference)
416
if unusual_modes != {}:
417
for path, mode in unusual_modes.iteritems():
418
warn_unusual_mode(rev.foreign_revid, path, mode)
419
mapping.import_unusual_file_modes(rev, unusual_modes)
421
basis_id = rev.parent_ids[0]
423
basis_id = NULL_REVISION
424
base_bzr_inventory = None
426
base_bzr_inventory = base_bzr_tree.root_inventory
427
rev.inventory_sha1, inv = repo.add_inventory_by_delta(
428
basis_id, inv_delta, rev.revision_id, rev.parent_ids,
430
ret_tree = InventoryRevisionTree(repo, inv, rev.revision_id)
432
if verifiers and roundtrip_revid is not None:
433
testament = StrictTestament3(rev, ret_tree)
434
calculated_verifiers = {"testament3-sha1": testament.as_sha1()}
435
if calculated_verifiers != verifiers:
436
trace.mutter("Testament SHA1 %r for %r did not match %r.",
437
calculated_verifiers["testament3-sha1"],
438
rev.revision_id, verifiers["testament3-sha1"])
439
rev.revision_id = original_revid
440
rev.inventory_sha1, inv = repo.add_inventory_by_delta(
441
basis_id, inv_delta, rev.revision_id, rev.parent_ids,
443
ret_tree = InventoryRevisionTree(repo, inv, rev.revision_id)
445
calculated_verifiers = {}
446
store_updater.add_object(o, calculated_verifiers, None)
447
store_updater.finish()
448
trees_cache.add(ret_tree)
449
repo.add_revision(rev.revision_id, rev)
450
if "verify" in debug.debug_flags:
451
verify_commit_reconstruction(
452
target_git_object_retriever, lookup_object, o, rev, ret_tree,
453
parent_trees, mapping, unusual_modes, verifiers)
456
def import_git_objects(repo, mapping, object_iter,
457
target_git_object_retriever, heads, pb=None,
459
"""Import a set of git objects into a bzr repository.
461
:param repo: Target Bazaar repository
462
:param mapping: Mapping to use
463
:param object_iter: Iterator over Git objects.
464
:return: Tuple with pack hints and last imported revision id
466
def lookup_object(sha):
468
return object_iter[sha]
470
return target_git_object_retriever[sha]
473
heads = list(set(heads))
474
trees_cache = LRUTreeCache(repo)
475
# Find and convert commit objects
478
pb.update("finding revisions to fetch", len(graph), None)
482
if not isinstance(head, bytes):
483
raise TypeError(head)
485
o = lookup_object(head)
488
if isinstance(o, Commit):
489
rev, roundtrip_revid, verifiers = mapping.import_commit(
490
o, mapping.revision_id_foreign_to_bzr, strict=True)
491
if (repo.has_revision(rev.revision_id)
492
or (roundtrip_revid and
493
repo.has_revision(roundtrip_revid))):
495
graph.append((o.id, o.parents))
496
heads.extend([p for p in o.parents if p not in checked])
497
elif isinstance(o, Tag):
498
if o.object[1] not in checked:
499
heads.append(o.object[1])
501
trace.warning("Unable to import head object %r" % o)
504
# Order the revisions
505
# Create the inventory objects
507
revision_ids = topo_sort(graph)
509
if limit is not None:
510
revision_ids = revision_ids[:limit]
512
for offset in range(0, len(revision_ids), batch_size):
513
target_git_object_retriever.start_write_group()
515
repo.start_write_group()
517
for i, head in enumerate(
518
revision_ids[offset:offset + batch_size]):
520
pb.update("fetching revisions", offset + i,
522
import_git_commit(repo, mapping, head, lookup_object,
523
target_git_object_retriever, trees_cache,
526
except BaseException:
527
repo.abort_write_group()
530
hint = repo.commit_write_group()
532
pack_hints.extend(hint)
533
except BaseException:
534
target_git_object_retriever.abort_write_group()
537
target_git_object_retriever.commit_write_group()
538
return pack_hints, last_imported
541
class DetermineWantsRecorder(object):
543
def __init__(self, actual):
546
self.remote_refs = {}
548
def __call__(self, refs):
549
if type(refs) is not dict:
550
raise TypeError(refs)
551
self.remote_refs = refs
552
self.wants = self.actual(refs)