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 cStringIO import (
21
from dulwich.client import (
22
SimpleFetchGraphWalker,
24
from dulwich.objects import (
35
from bzrlib.errors import (
39
from bzrlib.inventory import (
45
from bzrlib.lru_cache import (
48
from bzrlib.repository import (
51
from bzrlib.revision import (
54
from bzrlib.tsort import (
58
from bzrlib.plugins.git.converter import (
61
from bzrlib.plugins.git.repository import (
66
from bzrlib.plugins.git.remote import (
71
class BzrFetchGraphWalker(object):
72
"""GraphWalker implementation that uses a Bazaar repository."""
74
def __init__(self, repository, mapping):
75
self.repository = repository
76
self.mapping = mapping
78
self.heads = set(repository.all_revision_ids())
82
return iter(self.next, None)
85
revid = self.mapping.revision_id_foreign_to_bzr(sha)
88
def remove(self, revid):
90
if revid in self.heads:
91
self.heads.remove(revid)
92
if revid in self.parents:
93
for p in self.parents[revid]:
98
ret = self.heads.pop()
99
ps = self.repository.get_parent_map([ret])[ret]
100
self.parents[ret] = ps
101
self.heads.update([p for p in ps if not p in self.done])
104
return self.mapping.revision_id_bzr_to_foreign(ret)[0]
105
except InvalidRevisionId:
110
def import_git_blob(texts, mapping, path, hexsha, base_inv, parent_id,
111
revision_id, parent_invs, shagitmap, lookup_object, executable, symlink):
112
"""Import a git blob object into a bzr repository.
114
:param texts: VersionedFiles to add to
115
:param path: Path in the tree
116
:param blob: A git blob
117
:return: Inventory delta for this file
119
file_id = mapping.generate_file_id(path)
124
# We just have to hope this is indeed utf-8:
125
ie = cls(file_id, urlutils.basename(path).decode("utf-8"),
127
ie.executable = executable
128
# See if this has changed at all
130
base_sha = shagitmap.lookup_blob(file_id, base_inv.revision_id)
134
if (base_sha == hexsha and base_inv[file_id].executable == ie.executable
135
and base_inv[file_id].kind == ie.kind):
136
# If nothing has changed since the base revision, we're done
138
if base_sha == hexsha:
139
ie.text_size = base_inv[file_id].text_size
140
ie.text_sha1 = base_inv[file_id].text_sha1
141
ie.symlink_target = base_inv[file_id].symlink_target
142
ie.revision = base_inv[file_id].revision
144
blob = lookup_object(hexsha)
145
if ie.kind == "symlink":
146
ie.symlink_target = blob.data
150
ie.text_size = len(blob.data)
151
ie.text_sha1 = osutils.sha_string(blob.data)
152
# Check what revision we should store
154
for pinv in parent_invs:
155
if not file_id in pinv:
157
if pinv[file_id].text_sha1 == ie.text_sha1:
158
# found a revision in one of the parents to use
159
ie.revision = pinv[file_id].revision
161
parent_keys.append((file_id, pinv[file_id].revision))
162
if ie.revision is None:
163
# Need to store a new revision
164
ie.revision = revision_id
165
assert file_id is not None
166
assert ie.revision is not None
167
texts.add_lines((file_id, ie.revision), parent_keys,
168
osutils.split_lines(blob.data))
169
shagitmap.add_entry(hexsha, "blob", (ie.file_id, ie.revision))
170
if file_id in base_inv:
171
old_path = base_inv.id2path(file_id)
174
return [(old_path, path, file_id, ie)]
177
def import_git_tree(texts, mapping, path, hexsha, base_inv, parent_id,
178
revision_id, parent_invs, shagitmap, lookup_object):
179
"""Import a git tree object into a bzr repository.
181
:param texts: VersionedFiles object to add to
182
:param path: Path in the tree
183
:param tree: A git tree object
184
:param base_inv: Base inventory against which to return inventory delta
185
:return: Inventory delta for this subtree
188
file_id = mapping.generate_file_id(path)
189
# We just have to hope this is indeed utf-8:
190
ie = InventoryDirectory(file_id, urlutils.basename(path.decode("utf-8")),
192
if not file_id in base_inv:
193
# Newly appeared here
194
ie.revision = revision_id
195
texts.add_lines((file_id, ie.revision), [], [])
196
ret.append((None, path, file_id, ie))
198
# See if this has changed at all
200
base_sha = shagitmap.lookup_tree(path, base_inv.revision_id)
204
if base_sha == hexsha:
205
# If nothing has changed since the base revision, we're done
207
# Remember for next time
208
existing_children = set()
209
shagitmap.add_entry(hexsha, "tree", (file_id, revision_id))
210
tree = lookup_object(hexsha)
211
for mode, name, hexsha in tree.entries():
212
entry_kind = (mode & 0700000) / 0100000
213
basename = name.decode("utf-8")
214
existing_children.add(basename)
218
child_path = urlutils.join(path, name)
220
ret.extend(import_git_tree(texts, mapping, child_path, hexsha, base_inv,
221
file_id, revision_id, parent_invs, shagitmap, lookup_object))
222
elif entry_kind == 1:
223
fs_mode = mode & 0777
224
file_kind = (mode & 070000) / 010000
225
if file_kind == 0: # regular file
230
raise AssertionError("Unknown file kind, mode=%r" % (mode,))
231
ret.extend(import_git_blob(texts, mapping, child_path, hexsha, base_inv,
232
file_id, revision_id, parent_invs, shagitmap, lookup_object,
233
bool(fs_mode & 0111), symlink))
235
raise AssertionError("Unknown object kind, perms=%r." % (mode,))
236
# Remove any children that have disappeared
237
if file_id in base_inv:
238
deletable = [v for k,v in base_inv[file_id].children.iteritems() if k not in existing_children]
241
ret.append((base_inv.id2path(ie.file_id), None, ie.file_id, None))
242
if ie.kind == "directory":
243
deletable.extend(ie.children.values())
247
def import_git_objects(repo, mapping, object_iter, target_git_object_retriever,
249
"""Import a set of git objects into a bzr repository.
251
:param repo: Bazaar repository
252
:param mapping: Mapping to use
253
:param object_iter: Iterator over Git objects.
255
# TODO: a more (memory-)efficient implementation of this
261
parent_invs_cache = LRUCache(50)
262
# Find and convert commit objects
265
pb.update("finding revisions to fetch", len(graph), None)
267
assert isinstance(head, str)
269
o = object_iter[head]
272
if isinstance(o, Commit):
273
rev = mapping.import_commit(o)
274
if repo.has_revision(rev.revision_id):
276
root_trees[rev.revision_id] = o.tree
277
revisions[rev.revision_id] = rev
278
graph.append((rev.revision_id, rev.parent_ids))
279
target_git_object_retriever._idmap.add_entry(o.sha().hexdigest(),
280
"commit", (rev.revision_id, o._tree))
281
heads.extend([p for p in o.parents if p not in checked])
282
elif isinstance(o, Tag):
283
heads.append(o.object[1])
285
trace.warning("Unable to import head object %r" % o)
287
# Order the revisions
288
# Create the inventory objects
289
for i, revid in enumerate(topo_sort(graph)):
291
pb.update("fetching revisions", i, len(graph))
292
rev = revisions[revid]
293
# We have to do this here, since we have to walk the tree and
294
# we need to make sure to import the blobs / trees with the right
295
# path; this may involve adding them more than once.
296
def lookup_object(sha):
298
return object_iter[sha]
300
return target_git_object_retriever[sha]
302
for parent_id in rev.parent_ids:
304
parent_invs.append(parent_invs_cache[parent_id])
306
parent_inv = repo.get_inventory(parent_id)
307
parent_invs.append(parent_inv)
308
parent_invs_cache[parent_id] = parent_inv
309
if parent_invs == []:
310
base_inv = Inventory(root_id=None)
312
base_inv = parent_invs[0]
313
inv_delta = import_git_tree(repo.texts, mapping, "",
314
root_trees[revid], base_inv, None, revid, parent_invs,
315
target_git_object_retriever._idmap, lookup_object)
317
basis_id = rev.parent_ids[0]
319
basis_id = NULL_REVISION
320
rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,
321
inv_delta, rev.revision_id, rev.parent_ids)
322
parent_invs_cache[rev.revision_id] = inv
323
repo.add_revision(rev.revision_id, rev)
324
target_git_object_retriever._idmap.commit()
327
class InterGitNonGitRepository(InterRepository):
328
"""Base InterRepository that copies revisions from a Git into a non-Git
331
_matching_repo_format = GitRepositoryFormat()
334
def _get_repo_format_to_test():
337
def copy_content(self, revision_id=None, pb=None):
338
"""See InterRepository.copy_content."""
339
self.fetch(revision_id, pb, find_ghosts=False)
341
def fetch(self, revision_id=None, pb=None, find_ghosts=False, mapping=None,
343
self.fetch_refs(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts,
344
mapping=mapping, fetch_spec=fetch_spec)
346
def fetch_refs(self, revision_id=None, pb=None, find_ghosts=False,
347
mapping=None, fetch_spec=None):
349
mapping = self.source.get_mapping()
350
if revision_id is not None:
351
interesting_heads = [revision_id]
352
elif fetch_spec is not None:
353
interesting_heads = fetch_spec.heads
355
interesting_heads = None
357
def determine_wants(refs):
359
if interesting_heads is None:
360
ret = [sha for (ref, sha) in refs.iteritems() if not ref.endswith("^{}")]
362
ret = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in interesting_heads]
363
return [rev for rev in ret if not self.target.has_revision(mapping.revision_id_foreign_to_bzr(rev))]
364
self.fetch_objects(determine_wants, mapping, pb)
369
class InterRemoteGitNonGitRepository(InterGitNonGitRepository):
370
"""InterRepository that copies revisions from a remote Git into a non-Git
373
def fetch_objects(self, determine_wants, mapping, pb=None):
375
pb.update("git: %s" % text.rstrip("\r\n"), 0, 0)
376
graph_walker = BzrFetchGraphWalker(self.target, mapping)
379
create_pb = pb = ui.ui_factory.nested_progress_bar()
380
target_git_object_retriever = BazaarObjectStore(self.target, mapping)
383
def record_determine_wants(heads):
384
wants = determine_wants(heads)
385
recorded_wants.extend(wants)
389
self.target.lock_write()
391
self.target.start_write_group()
393
objects_iter = self.source.fetch_objects(
394
record_determine_wants,
396
target_git_object_retriever.get_raw,
398
import_git_objects(self.target, mapping, objects_iter,
399
target_git_object_retriever, recorded_wants, pb)
401
self.target.commit_write_group()
409
def is_compatible(source, target):
410
"""Be compatible with GitRepository."""
411
# FIXME: Also check target uses VersionedFile
412
return (isinstance(source, RemoteGitRepository) and
413
target.supports_rich_root() and
414
not isinstance(target, GitRepository))
417
class InterLocalGitNonGitRepository(InterGitNonGitRepository):
418
"""InterRepository that copies revisions from a remote Git into a non-Git
421
def fetch_objects(self, determine_wants, mapping, pb=None):
422
wants = determine_wants(self.source._git.get_refs())
425
create_pb = pb = ui.ui_factory.nested_progress_bar()
426
target_git_object_retriever = BazaarObjectStore(self.target, mapping)
428
self.target.lock_write()
430
self.target.start_write_group()
432
import_git_objects(self.target, mapping,
433
self.source._git.object_store,
434
target_git_object_retriever, wants, pb)
436
self.target.commit_write_group()
444
def is_compatible(source, target):
445
"""Be compatible with GitRepository."""
446
# FIXME: Also check target uses VersionedFile
447
return (isinstance(source, LocalGitRepository) and
448
target.supports_rich_root() and
449
not isinstance(target, GitRepository))
452
class InterGitRepository(InterRepository):
453
"""InterRepository that copies between Git repositories."""
455
_matching_repo_format = GitRepositoryFormat()
458
def _get_repo_format_to_test():
461
def copy_content(self, revision_id=None, pb=None):
462
"""See InterRepository.copy_content."""
463
self.fetch(revision_id, pb, find_ghosts=False)
465
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
466
mapping=None, fetch_spec=None):
468
mapping = self.source.get_mapping()
470
trace.info("git: %s", text)
472
if revision_id is not None:
473
args = [mapping.revision_id_bzr_to_foreign(revision_id)[0]]
474
elif fetch_spec is not None:
475
args = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in fetch_spec.heads]
476
if fetch_spec is None and revision_id is None:
477
determine_wants = r.object_store.determine_wants_all
479
determine_wants = lambda x: [y for y in args if not y in r.object_store]
481
graphwalker = SimpleFetchGraphWalker(r.heads().values(), r.get_parents)
482
f, commit = r.object_store.add_pack()
484
self.source._git.fetch_pack(path, determine_wants, graphwalker, f.write, progress)
492
def is_compatible(source, target):
493
"""Be compatible with GitRepository."""
494
return (isinstance(source, GitRepository) and
495
isinstance(target, GitRepository))
added
removed
fetch.py
