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 (
32
from dulwich.walk import Walker
43
from ..errors import (
46
from ..bzr.inventory import (
52
from ..repository import (
55
from ..revision import (
58
from ..bzr.inventorytree import InventoryRevisionTree
59
from ..sixish import text_type
60
from ..testament import (
66
from ..bzr.versionedfile import (
67
ChunkedContentFactory,
70
from .mapping import (
76
from .object_store import (
86
from .repository import (
93
def import_git_blob(texts, mapping, path, name, hexshas,
94
base_bzr_tree, parent_id, revision_id,
95
parent_bzr_trees, lookup_object, modes, 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 not isinstance(path, bytes):
105
raise TypeError(path)
106
decoded_path = path.decode('utf-8')
107
(base_mode, mode) = modes
108
(base_hexsha, hexsha) = hexshas
109
if mapping.is_special_file(path):
111
if base_hexsha == hexsha and base_mode == mode:
112
# If nothing has changed since the base revision, we're done
114
file_id = lookup_file_id(decoded_path)
115
if stat.S_ISLNK(mode):
119
ie = cls(file_id, name.decode("utf-8"), parent_id)
120
if ie.kind == "file":
121
ie.executable = mode_is_executable(mode)
122
if base_hexsha == hexsha and mode_kind(base_mode) == mode_kind(mode):
123
base_exec = base_bzr_tree.is_executable(decoded_path)
124
if ie.kind == "symlink":
125
ie.symlink_target = base_bzr_tree.get_symlink_target(decoded_path)
127
ie.text_size = base_bzr_tree.get_file_size(decoded_path)
128
ie.text_sha1 = base_bzr_tree.get_file_sha1(decoded_path)
129
if ie.kind == "symlink" or ie.executable == base_exec:
130
ie.revision = base_bzr_tree.get_file_revision(decoded_path)
132
blob = lookup_object(hexsha)
134
blob = lookup_object(hexsha)
135
if ie.kind == "symlink":
137
ie.symlink_target = blob.data.decode("utf-8")
139
ie.text_size = sum(map(len, blob.chunked))
140
ie.text_sha1 = osutils.sha_strings(blob.chunked)
141
# Check what revision we should store
143
for ptree in parent_bzr_trees:
145
ppath = ptree.id2path(file_id)
146
except errors.NoSuchId:
148
pkind = ptree.kind(ppath, file_id)
149
if (pkind == ie.kind and
150
((pkind == "symlink" and ptree.get_symlink_target(ppath, file_id) == ie.symlink_target) or
151
(pkind == "file" and ptree.get_file_sha1(ppath, file_id) == ie.text_sha1 and
152
ptree.is_executable(ppath, file_id) == ie.executable))):
153
# found a revision in one of the parents to use
154
ie.revision = ptree.get_file_revision(ppath, file_id)
156
parent_key = (file_id, ptree.get_file_revision(ppath, file_id))
157
if not parent_key in parent_keys:
158
parent_keys.append(parent_key)
159
if ie.revision is None:
160
# Need to store a new revision
161
ie.revision = revision_id
162
if ie.revision is None:
163
raise ValueError("no file revision set")
164
if ie.kind == 'symlink':
167
chunks = blob.chunked
168
texts.insert_record_stream([
169
ChunkedContentFactory((file_id, ie.revision),
170
tuple(parent_keys), ie.text_sha1, chunks)])
172
if base_hexsha is not None:
173
old_path = decoded_path # Renames are not supported yet
174
if stat.S_ISDIR(base_mode):
175
invdelta.extend(remove_disappeared_children(base_bzr_tree, old_path,
176
lookup_object(base_hexsha), [], lookup_object))
179
invdelta.append((old_path, decoded_path, file_id, ie))
180
if base_hexsha != hexsha:
181
store_updater.add_object(blob, (ie.file_id, ie.revision), path)
185
class SubmodulesRequireSubtrees(BzrError):
186
_fmt = ("The repository you are fetching from contains submodules, "
187
"which require a Bazaar format that supports tree references.")
191
def import_git_submodule(texts, mapping, path, name, hexshas,
192
base_bzr_tree, parent_id, revision_id, parent_bzr_trees, lookup_object,
193
modes, store_updater, lookup_file_id):
194
"""Import a git submodule."""
195
(base_hexsha, hexsha) = hexshas
196
(base_mode, mode) = modes
197
if base_hexsha == hexsha and base_mode == mode:
199
file_id = lookup_file_id(path)
201
ie = TreeReference(file_id, name.decode("utf-8"), parent_id)
202
ie.revision = revision_id
203
if base_hexsha is not None:
204
old_path = path.decode("utf-8") # Renames are not supported yet
205
if stat.S_ISDIR(base_mode):
206
invdelta.extend(remove_disappeared_children(base_bzr_tree, old_path,
207
lookup_object(base_hexsha), [], lookup_object))
210
ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)
211
texts.insert_record_stream([
212
ChunkedContentFactory((file_id, ie.revision), (), None, [])])
213
invdelta.append((old_path, path, file_id, ie))
217
def remove_disappeared_children(base_bzr_tree, path, base_tree, existing_children,
219
"""Generate an inventory delta for removed children.
221
:param base_bzr_tree: Base bzr tree against which to generate the
223
:param path: Path to process (unicode)
224
:param base_tree: Git Tree base object
225
:param existing_children: Children that still exist
226
:param lookup_object: Lookup a git object by its SHA1
227
:return: Inventory delta, as list
229
if not isinstance(path, text_type):
230
raise TypeError(path)
232
for name, mode, hexsha in base_tree.iteritems():
233
if name in existing_children:
235
c_path = posixpath.join(path, name.decode("utf-8"))
236
file_id = base_bzr_tree.path2id(c_path)
238
raise TypeError(file_id)
239
ret.append((c_path, None, file_id, None))
240
if stat.S_ISDIR(mode):
241
ret.extend(remove_disappeared_children(
242
base_bzr_tree, c_path, lookup_object(hexsha), [], lookup_object))
246
def import_git_tree(texts, mapping, path, name, hexshas,
247
base_bzr_tree, parent_id, revision_id, parent_bzr_trees,
248
lookup_object, modes, store_updater,
249
lookup_file_id, allow_submodules=False):
250
"""Import a git tree object into a bzr repository.
252
:param texts: VersionedFiles object to add to
253
:param path: Path in the tree (str)
254
:param name: Name of the tree (str)
255
:param tree: A git tree object
256
:param base_bzr_tree: Base inventory against which to return inventory delta
257
:return: Inventory delta for this subtree
259
(base_hexsha, hexsha) = hexshas
260
(base_mode, mode) = modes
261
if not isinstance(path, bytes):
262
raise TypeError(path)
263
if not isinstance(name, bytes):
264
raise TypeError(name)
265
if base_hexsha == hexsha and base_mode == mode:
266
# If nothing has changed since the base revision, we're done
269
file_id = lookup_file_id(osutils.safe_unicode(path))
270
# We just have to hope this is indeed utf-8:
271
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
272
tree = lookup_object(hexsha)
273
if base_hexsha is None:
275
old_path = None # Newly appeared here
277
base_tree = lookup_object(base_hexsha)
278
old_path = path.decode("utf-8") # Renames aren't supported yet
279
new_path = path.decode("utf-8")
280
if base_tree is None or type(base_tree) is not Tree:
281
ie.revision = revision_id
282
invdelta.append((old_path, new_path, ie.file_id, ie))
283
texts.insert_record_stream([
284
ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
285
# Remember for next time
286
existing_children = set()
288
for name, child_mode, child_hexsha in tree.iteritems():
289
existing_children.add(name)
290
child_path = posixpath.join(path, name)
291
if type(base_tree) is Tree:
293
child_base_mode, child_base_hexsha = base_tree[name]
295
child_base_hexsha = None
298
child_base_hexsha = None
300
if stat.S_ISDIR(child_mode):
301
subinvdelta, grandchildmodes = import_git_tree(texts, mapping,
302
child_path, name, (child_base_hexsha, child_hexsha),
303
base_bzr_tree, file_id, revision_id, parent_bzr_trees,
304
lookup_object, (child_base_mode, child_mode), store_updater,
305
lookup_file_id, allow_submodules=allow_submodules)
306
elif S_ISGITLINK(child_mode): # submodule
307
if not allow_submodules:
308
raise SubmodulesRequireSubtrees()
309
subinvdelta, grandchildmodes = import_git_submodule(texts, mapping,
310
child_path, name, (child_base_hexsha, child_hexsha),
311
base_bzr_tree, file_id, revision_id, parent_bzr_trees,
312
lookup_object, (child_base_mode, child_mode), store_updater,
315
if not mapping.is_special_file(name):
316
subinvdelta = import_git_blob(texts, mapping, child_path, name,
317
(child_base_hexsha, child_hexsha), base_bzr_tree, file_id,
318
revision_id, parent_bzr_trees, lookup_object,
319
(child_base_mode, child_mode), store_updater, lookup_file_id)
323
child_modes.update(grandchildmodes)
324
invdelta.extend(subinvdelta)
325
if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
326
stat.S_IFLNK, DEFAULT_FILE_MODE|0o111,
328
child_modes[child_path] = child_mode
329
# Remove any children that have disappeared
330
if base_tree is not None and type(base_tree) is Tree:
331
invdelta.extend(remove_disappeared_children(base_bzr_tree, old_path,
332
base_tree, existing_children, lookup_object))
333
store_updater.add_object(tree, (file_id, revision_id), path)
334
return invdelta, child_modes
337
def verify_commit_reconstruction(target_git_object_retriever, lookup_object,
338
o, rev, ret_tree, parent_trees, mapping, 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(ret_tree, parent_trees,
352
target_git_object_retriever._cache.idmap, unusual_modes,
353
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" and
360
new_obj.type_name == "tree" and
361
sorted(old_obj) == sorted(new_obj)):
363
if old_obj[name][0] != new_obj[name][0]:
364
raise AssertionError("Modes for %s differ: %o != %o" %
365
(path, old_obj[name][0], new_obj[name][0]))
366
if old_obj[name][1] != new_obj[name][1]:
367
# Found a differing child, delve deeper
368
path = posixpath.join(path, name)
369
old_obj = lookup_object(old_obj[name][1])
370
new_obj = new_objs[path]
372
raise AssertionError("objects differ for %s: %r != %r" % (path,
376
def ensure_inventories_in_repo(repo, trees):
377
real_inv_vf = repo.inventories.without_fallbacks()
379
revid = t.get_revision_id()
380
if not real_inv_vf.get_parent_map([(revid, )]):
381
repo.add_inventory(revid, t.root_inventory, t.get_parent_ids())
384
def import_git_commit(repo, mapping, head, lookup_object,
385
target_git_object_retriever, trees_cache):
386
o = lookup_object(head)
387
# Note that this uses mapping.revision_id_foreign_to_bzr. If the parents
388
# were bzr roundtripped revisions they would be specified in the
390
rev, roundtrip_revid, verifiers = mapping.import_commit(
391
o, mapping.revision_id_foreign_to_bzr)
392
if roundtrip_revid is not None:
393
original_revid = rev.revision_id
394
rev.revision_id = roundtrip_revid
395
# We have to do this here, since we have to walk the tree and
396
# we need to make sure to import the blobs / trees with the right
397
# path; this may involve adding them more than once.
398
parent_trees = trees_cache.revision_trees(rev.parent_ids)
399
ensure_inventories_in_repo(repo, parent_trees)
400
if parent_trees == []:
401
base_bzr_tree = trees_cache.revision_tree(NULL_REVISION)
405
base_bzr_tree = parent_trees[0]
406
base_tree = lookup_object(o.parents[0]).tree
407
base_mode = stat.S_IFDIR
408
store_updater = target_git_object_retriever._get_updater(rev)
409
tree_supplement = mapping.get_fileid_map(lookup_object, o.tree)
410
inv_delta, unusual_modes = import_git_tree(repo.texts,
411
mapping, b"", b"", (base_tree, o.tree), base_bzr_tree,
412
None, rev.revision_id, parent_trees,
413
lookup_object, (base_mode, stat.S_IFDIR), store_updater,
414
tree_supplement.lookup_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(basis_id,
428
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(basis_id,
441
inv_delta, rev.revision_id, rev.parent_ids, base_bzr_tree)
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(target_git_object_retriever,
451
lookup_object, o, rev, ret_tree, parent_trees, mapping,
452
unusual_modes, verifiers)
455
def import_git_objects(repo, mapping, object_iter,
456
target_git_object_retriever, heads, pb=None, limit=None):
457
"""Import a set of git objects into a bzr repository.
459
:param repo: Target Bazaar repository
460
:param mapping: Mapping to use
461
:param object_iter: Iterator over Git objects.
462
:return: Tuple with pack hints and last imported revision id
464
def lookup_object(sha):
466
return object_iter[sha]
468
return target_git_object_retriever[sha]
471
heads = list(set(heads))
472
trees_cache = LRUTreeCache(repo)
473
# Find and convert commit objects
476
pb.update("finding revisions to fetch", len(graph), None)
480
if not isinstance(head, bytes):
481
raise TypeError(head)
483
o = lookup_object(head)
486
if isinstance(o, Commit):
487
rev, roundtrip_revid, verifiers = mapping.import_commit(o,
488
mapping.revision_id_foreign_to_bzr)
489
if (repo.has_revision(rev.revision_id) or
490
(roundtrip_revid and repo.has_revision(roundtrip_revid))):
492
graph.append((o.id, o.parents))
493
heads.extend([p for p in o.parents if p not in checked])
494
elif isinstance(o, Tag):
495
if o.object[1] not in checked:
496
heads.append(o.object[1])
498
trace.warning("Unable to import head object %r" % o)
501
# Order the revisions
502
# Create the inventory objects
504
revision_ids = topo_sort(graph)
506
if limit is not None:
507
revision_ids = revision_ids[:limit]
509
for offset in range(0, len(revision_ids), batch_size):
510
target_git_object_retriever.start_write_group()
512
repo.start_write_group()
514
for i, head in enumerate(
515
revision_ids[offset:offset+batch_size]):
517
pb.update("fetching revisions", offset+i,
519
import_git_commit(repo, mapping, head, lookup_object,
520
target_git_object_retriever, trees_cache)
523
repo.abort_write_group()
526
hint = repo.commit_write_group()
528
pack_hints.extend(hint)
530
target_git_object_retriever.abort_write_group()
533
target_git_object_retriever.commit_write_group()
534
return pack_hints, last_imported
537
class DetermineWantsRecorder(object):
539
def __init__(self, actual):
542
self.remote_refs = {}
544
def __call__(self, refs):
545
if type(refs) is not dict:
546
raise TypeError(refs)
547
self.remote_refs = refs
548
self.wants = self.actual(refs)