1
# Copyright (C) 2009-2018 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
"""InterRepository operations."""
19
from __future__ import absolute_import
21
from io import BytesIO
23
from dulwich.errors import (
26
from dulwich.object_store import (
27
ObjectStoreGraphWalker,
29
from dulwich.protocol import (
33
from dulwich.walk import Walker
35
from ...errors import (
36
FetchLimitUnsupported,
39
NoRoundtrippingSupport,
42
from ...repository import (
45
from ...revision import (
59
DetermineWantsRecorder,
61
from .mapping import (
64
from .object_store import (
69
MissingObjectsIterator,
74
from .repository import (
82
from .unpeel_map import (
87
class InterToGitRepository(InterRepository):
88
"""InterRepository that copies into a Git repository."""
90
_matching_repo_format = GitRepositoryFormat()
92
def __init__(self, source, target):
93
super(InterToGitRepository, self).__init__(source, target)
94
self.mapping = self.target.get_mapping()
95
self.source_store = get_object_store(self.source, self.mapping)
98
def _get_repo_format_to_test():
101
def copy_content(self, revision_id=None, pb=None):
102
"""See InterRepository.copy_content."""
103
self.fetch(revision_id, pb, find_ghosts=False)
105
def fetch_refs(self, update_refs, lossy):
106
"""Fetch possibly roundtripped revisions into the target repository
109
:param update_refs: Generate refs to fetch. Receives dictionary
110
with old refs (git shas), returns dictionary of new names to
112
:param lossy: Whether to roundtrip
113
:return: old refs, new refs
115
raise NotImplementedError(self.fetch_refs)
117
def search_missing_revision_ids(self,
118
find_ghosts=True, revision_ids=None, if_present_ids=None,
120
if limit is not None:
121
raise FetchLimitUnsupported(self)
125
todo.extend(revision_ids)
127
todo.extend(revision_ids)
128
with self.source_store.lock_read():
129
for revid in revision_ids:
130
if revid == NULL_REVISION:
132
git_sha = self.source_store._lookup_revision_sha1(revid)
133
git_shas.append(git_sha)
134
walker = Walker(self.source_store,
135
include=git_shas, exclude=[sha for sha in self.target.controldir.get_refs_container().as_dict().values() if sha != ZERO_SHA])
136
missing_revids = set()
138
for (kind, type_data) in self.source_store.lookup_git_sha(entry.commit.id):
140
missing_revids.add(type_data[0])
141
return self.source.revision_ids_to_search_result(missing_revids)
143
def _warn_slow(self):
145
'Pushing from a Bazaar to a Git repository. '
146
'For better performance, push into a Bazaar repository.')
149
class InterToLocalGitRepository(InterToGitRepository):
150
"""InterBranch implementation between a Bazaar and a Git repository."""
152
def __init__(self, source, target):
153
super(InterToLocalGitRepository, self).__init__(source, target)
154
self.target_store = self.target.controldir._git.object_store
155
self.target_refs = self.target.controldir._git.refs
157
def _commit_needs_fetching(self, sha_id):
159
return (sha_id not in self.target_store)
160
except NoSuchRevision:
164
def _revision_needs_fetching(self, sha_id, revid):
165
if revid == NULL_REVISION:
169
sha_id = self.source_store._lookup_revision_sha1(revid)
172
return self._commit_needs_fetching(sha_id)
174
def missing_revisions(self, stop_revisions):
175
"""Find the revisions that are missing from the target repository.
177
:param stop_revisions: Revisions to check for (tuples with
179
:return: sequence of missing revisions, in topological order
180
:raise: NoSuchRevision if the stop_revisions are not present in
185
for (sha1, revid) in stop_revisions:
186
if sha1 is not None and revid is not None:
187
revid_sha_map[revid] = sha1
188
stop_revids.append(revid)
189
elif sha1 is not None:
190
if self._commit_needs_fetching(sha1):
191
for (kind, (revid, tree_sha, verifiers)) in self.source_store.lookup_git_sha(sha1):
192
revid_sha_map[revid] = sha1
193
stop_revids.append(revid)
197
stop_revids.append(revid)
199
graph = self.source.get_graph()
200
pb = ui.ui_factory.nested_progress_bar()
204
for revid in stop_revids:
205
sha1 = revid_sha_map.get(revid)
206
if (not revid in missing and
207
self._revision_needs_fetching(sha1, revid)):
209
new_stop_revids.append(revid)
211
parent_map = graph.get_parent_map(new_stop_revids)
212
for parent_revids in parent_map.itervalues():
213
stop_revids.update(parent_revids)
214
pb.update("determining revisions to fetch", len(missing))
217
return graph.iter_topo_order(missing)
219
def _get_target_bzr_refs(self):
220
"""Return a dictionary with references.
222
:return: Dictionary with reference names as keys and tuples
223
with Git SHA, Bazaar revid as values.
227
for k in self.target._git.refs.allkeys():
229
v = self.target._git.refs[k]
234
for (kind, type_data) in self.source_store.lookup_git_sha(v):
235
if kind == "commit" and self.source.has_revision(type_data[0]):
242
bzr_refs[k] = (v, revid)
245
def fetch_refs(self, update_refs, lossy):
246
with self.source_store.lock_read():
247
old_refs = self._get_target_bzr_refs()
248
new_refs = update_refs(old_refs)
249
revidmap = self.fetch_objects(
250
[(git_sha, bzr_revid) for (git_sha, bzr_revid) in new_refs.values() if git_sha is None or not git_sha.startswith('ref:')], lossy=lossy)
251
for name, (gitid, revid) in new_refs.iteritems():
254
gitid = revidmap[revid][0]
256
gitid = self.source_store._lookup_revision_sha1(revid)
257
if len(gitid) != 40 and not gitid.startswith('ref: '):
258
raise AssertionError("invalid ref contents: %r" % gitid)
259
self.target_refs[name] = gitid
260
return revidmap, old_refs, new_refs
262
def fetch_objects(self, revs, lossy, limit=None):
263
if not lossy and not self.mapping.roundtripping:
264
for git_sha, bzr_revid in revs:
265
if bzr_revid is not None and needs_roundtripping(self.source, bzr_revid):
266
raise NoPushSupport(self.source, self.target, self.mapping,
268
with self.source_store.lock_read():
269
todo = list(self.missing_revisions(revs))[:limit]
271
pb = ui.ui_factory.nested_progress_bar()
273
object_generator = MissingObjectsIterator(
274
self.source_store, self.source, pb)
275
for (old_revid, git_sha) in object_generator.import_revisions(
278
new_revid = self.mapping.revision_id_foreign_to_bzr(git_sha)
280
new_revid = old_revid
282
self.mapping.revision_id_bzr_to_foreign(old_revid)
283
except InvalidRevisionId:
284
refname = self.mapping.revid_as_refname(old_revid)
285
self.target_refs[refname] = git_sha
286
revidmap[old_revid] = (git_sha, new_revid)
287
self.target_store.add_objects(object_generator)
292
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
293
fetch_spec=None, mapped_refs=None):
294
if mapped_refs is not None:
295
stop_revisions = mapped_refs
296
elif revision_id is not None:
297
stop_revisions = [(None, revision_id)]
298
elif fetch_spec is not None:
299
recipe = fetch_spec.get_recipe()
300
if recipe[0] in ("search", "proxy-search"):
301
stop_revisions = [(None, revid) for revid in recipe[1]]
303
raise AssertionError("Unsupported search result type %s" % recipe[0])
305
stop_revisions = [(None, revid) for revid in self.source.all_revision_ids()]
308
self.fetch_objects(stop_revisions, lossy=False)
309
except NoPushSupport:
310
raise NoRoundtrippingSupport(self.source, self.target)
313
def is_compatible(source, target):
314
"""Be compatible with GitRepository."""
315
return (not isinstance(source, GitRepository) and
316
isinstance(target, LocalGitRepository))
319
class InterToRemoteGitRepository(InterToGitRepository):
321
def fetch_refs(self, update_refs, lossy):
322
"""Import the gist of the ancestry of a particular revision."""
323
if not lossy and not self.mapping.roundtripping:
324
raise NoPushSupport(self.source, self.target, self.mapping)
325
unpeel_map = UnpeelMap.from_repository(self.source)
327
def determine_wants(old_refs):
329
self.old_refs = dict([(k, (v, None)) for (k, v) in old_refs.iteritems()])
330
self.new_refs = update_refs(self.old_refs)
331
for name, (gitid, revid) in self.new_refs.iteritems():
333
git_sha = self.source_store._lookup_revision_sha1(revid)
334
ret[name] = unpeel_map.re_unpeel_tag(git_sha, old_refs.get(name))
339
with self.source_store.lock_read():
340
new_refs = self.target.send_pack(determine_wants,
341
self.source_store.generate_lossy_pack_data)
343
return revidmap, self.old_refs, self.new_refs
346
def is_compatible(source, target):
347
"""Be compatible with GitRepository."""
348
return (not isinstance(source, GitRepository) and
349
isinstance(target, RemoteGitRepository))
352
class InterFromGitRepository(InterRepository):
354
_matching_repo_format = GitRepositoryFormat()
356
def _target_has_shas(self, shas):
357
raise NotImplementedError(self._target_has_shas)
359
def get_determine_wants_heads(self, wants, include_tags=False):
361
def determine_wants(refs):
362
potential = set(wants)
364
for k, unpeeled in refs.iteritems():
365
if k.endswith("^{}"):
369
if unpeeled == ZERO_SHA:
371
potential.add(unpeeled)
372
return list(potential - self._target_has_shas(potential))
373
return determine_wants
375
def determine_wants_all(self, refs):
376
raise NotImplementedError(self.determine_wants_all)
379
def _get_repo_format_to_test():
382
def copy_content(self, revision_id=None):
383
"""See InterRepository.copy_content."""
384
self.fetch(revision_id, find_ghosts=False)
386
def search_missing_revision_ids(self,
387
find_ghosts=True, revision_ids=None, if_present_ids=None,
389
if limit is not None:
390
raise FetchLimitUnsupported(self)
394
todo.extend(revision_ids)
396
todo.extend(revision_ids)
397
for revid in revision_ids:
398
if revid == NULL_REVISION:
400
git_sha, mapping = self.source.lookup_bzr_revision_id(revid)
401
git_shas.append(git_sha)
402
walker = Walker(self.source._git.object_store,
403
include=git_shas, exclude=[
404
sha for sha in self.target.controldir.get_refs_container().as_dict().values()
406
missing_revids = set()
408
missing_revids.add(self.source.lookup_foreign_revision_id(entry.commit.id))
409
return self.source.revision_ids_to_search_result(missing_revids)
412
class InterGitNonGitRepository(InterFromGitRepository):
413
"""Base InterRepository that copies revisions from a Git into a non-Git
416
def _target_has_shas(self, shas):
420
revid = self.source.lookup_foreign_revision_id(sha)
421
except NotCommitError:
422
# Commit is definitely not present
426
return set([revids[r] for r in self.target.has_revisions(revids)])
428
def determine_wants_all(self, refs):
430
for k, v in refs.iteritems():
431
# For non-git target repositories, only worry about peeled
434
potential.add(self.source.controldir.get_peeled(k) or v)
435
return list(potential - self._target_has_shas(potential))
437
def get_determine_wants_heads(self, wants, include_tags=False):
439
def determine_wants(refs):
440
potential = set(wants)
442
for k, unpeeled in refs.iteritems():
445
if unpeeled == ZERO_SHA:
447
potential.add(self.source.controldir.get_peeled(k) or unpeeled)
448
return list(potential - self._target_has_shas(potential))
449
return determine_wants
451
def _warn_slow(self):
453
'Fetching from Git to Bazaar repository. '
454
'For better performance, fetch into a Git repository.')
456
def fetch_objects(self, determine_wants, mapping, limit=None, lossy=False):
457
"""Fetch objects from a remote server.
459
:param determine_wants: determine_wants callback
460
:param mapping: BzrGitMapping to use
461
:param limit: Maximum number of commits to import.
462
:return: Tuple with pack hint, last imported revision id and remote refs
464
raise NotImplementedError(self.fetch_objects)
466
def get_determine_wants_revids(self, revids, include_tags=False):
468
for revid in set(revids):
469
if self.target.has_revision(revid):
471
git_sha, mapping = self.source.lookup_bzr_revision_id(revid)
473
return self.get_determine_wants_heads(wants, include_tags=include_tags)
475
def fetch(self, revision_id=None, find_ghosts=False,
476
mapping=None, fetch_spec=None, include_tags=False):
478
mapping = self.source.get_mapping()
479
if revision_id is not None:
480
interesting_heads = [revision_id]
481
elif fetch_spec is not None:
482
recipe = fetch_spec.get_recipe()
483
if recipe[0] in ("search", "proxy-search"):
484
interesting_heads = recipe[1]
486
raise AssertionError("Unsupported search result type %s" %
489
interesting_heads = None
491
if interesting_heads is not None:
492
determine_wants = self.get_determine_wants_revids(
493
interesting_heads, include_tags=include_tags)
495
determine_wants = self.determine_wants_all
497
(pack_hint, _, remote_refs) = self.fetch_objects(determine_wants,
499
if pack_hint is not None and self.target._format.pack_compresses:
500
self.target.pack(hint=pack_hint)
504
class InterRemoteGitNonGitRepository(InterGitNonGitRepository):
505
"""InterRepository that copies revisions from a remote Git into a non-Git
508
def get_target_heads(self):
509
# FIXME: This should be more efficient
510
all_revs = self.target.all_revision_ids()
511
parent_map = self.target.get_parent_map(all_revs)
513
map(all_parents.update, parent_map.itervalues())
514
return set(all_revs) - all_parents
516
def fetch_objects(self, determine_wants, mapping, limit=None, lossy=False):
517
"""See `InterGitNonGitRepository`."""
519
store = get_object_store(self.target, mapping)
520
with store.lock_write():
521
heads = self.get_target_heads()
522
graph_walker = ObjectStoreGraphWalker(
523
[store._lookup_revision_sha1(head) for head in heads],
524
lambda sha: store[sha].parents)
525
wants_recorder = DetermineWantsRecorder(determine_wants)
527
pb = ui.ui_factory.nested_progress_bar()
529
objects_iter = self.source.fetch_objects(
530
wants_recorder, graph_walker, store.get_raw,
531
progress=lambda text: report_git_progress(pb, text),)
532
trace.mutter("Importing %d new revisions",
533
len(wants_recorder.wants))
534
(pack_hint, last_rev) = import_git_objects(self.target,
535
mapping, objects_iter, store, wants_recorder.wants, pb,
537
return (pack_hint, last_rev, wants_recorder.remote_refs)
542
def is_compatible(source, target):
543
"""Be compatible with GitRepository."""
544
if not isinstance(source, RemoteGitRepository):
546
if not target.supports_rich_root():
548
if isinstance(target, GitRepository):
550
if not getattr(target._format, "supports_full_versioned_files", True):
555
class InterLocalGitNonGitRepository(InterGitNonGitRepository):
556
"""InterRepository that copies revisions from a local Git into a non-Git
559
def fetch_objects(self, determine_wants, mapping, limit=None, lossy=False):
560
"""See `InterGitNonGitRepository`."""
562
remote_refs = self.source.controldir.get_refs_container().as_dict()
563
wants = determine_wants(remote_refs)
565
pb = ui.ui_factory.nested_progress_bar()
566
target_git_object_retriever = get_object_store(self.target, mapping)
568
target_git_object_retriever.lock_write()
570
(pack_hint, last_rev) = import_git_objects(self.target,
571
mapping, self.source._git.object_store,
572
target_git_object_retriever, wants, pb, limit)
573
return (pack_hint, last_rev, remote_refs)
575
target_git_object_retriever.unlock()
580
def is_compatible(source, target):
581
"""Be compatible with GitRepository."""
582
if not isinstance(source, LocalGitRepository):
584
if not target.supports_rich_root():
586
if isinstance(target, GitRepository):
588
if not getattr(target._format, "supports_full_versioned_files", True):
593
class InterGitGitRepository(InterFromGitRepository):
594
"""InterRepository that copies between Git repositories."""
596
def fetch_refs(self, update_refs, lossy=False):
598
raise LossyPushToSameVCS(self.source, self.target)
599
old_refs = self.target.controldir.get_refs_container()
601
def determine_wants(heads):
602
old_refs = dict([(k, (v, None)) for (k, v) in heads.as_dict().iteritems()])
603
new_refs = update_refs(old_refs)
604
ref_changes.update(new_refs)
605
return [sha1 for (sha1, bzr_revid) in new_refs.itervalues()]
606
self.fetch_objects(determine_wants, lossy=lossy)
607
for k, (git_sha, bzr_revid) in ref_changes.iteritems():
608
self.target._git.refs[k] = git_sha
609
new_refs = self.target.controldir.get_refs_container()
610
return None, old_refs, new_refs
612
def fetch_objects(self, determine_wants, mapping=None, limit=None, lossy=False):
613
raise NotImplementedError(self.fetch_objects)
615
def _target_has_shas(self, shas):
616
return set([sha for sha in shas if sha in self.target._git.object_store])
618
def fetch(self, revision_id=None, find_ghosts=False,
619
mapping=None, fetch_spec=None, branches=None, limit=None, include_tags=False):
621
mapping = self.source.get_mapping()
622
if revision_id is not None:
624
elif fetch_spec is not None:
625
recipe = fetch_spec.get_recipe()
626
if recipe[0] in ("search", "proxy-search"):
629
raise AssertionError(
630
"Unsupported search result type %s" % recipe[0])
632
if branches is not None:
633
def determine_wants(refs):
635
for name, value in refs.iteritems():
636
if value == ZERO_SHA:
639
if name in branches or (include_tags and is_tag(name)):
642
elif fetch_spec is None and revision_id is None:
643
determine_wants = self.determine_wants_all
645
determine_wants = self.get_determine_wants_revids(args, include_tags=include_tags)
646
wants_recorder = DetermineWantsRecorder(determine_wants)
647
self.fetch_objects(wants_recorder, mapping, limit=limit)
648
return wants_recorder.remote_refs
650
def get_determine_wants_revids(self, revids, include_tags=False):
652
for revid in set(revids):
653
if revid == NULL_REVISION:
655
git_sha, mapping = self.source.lookup_bzr_revision_id(revid)
657
return self.get_determine_wants_heads(wants, include_tags=include_tags)
659
def determine_wants_all(self, refs):
660
potential = set([v for v in refs.values() if not v == ZERO_SHA])
661
return list(potential - self._target_has_shas(potential))
664
class InterLocalGitLocalGitRepository(InterGitGitRepository):
666
def fetch_objects(self, determine_wants, mapping=None, limit=None, lossy=False):
668
raise LossyPushToSameVCS(self.source, self.target)
669
if limit is not None:
670
raise FetchLimitUnsupported(self)
671
pb = ui.ui_factory.nested_progress_bar()
673
refs = self.source._git.fetch(self.target._git, determine_wants,
674
lambda text: report_git_progress(pb, text))
677
return (None, None, refs)
680
def is_compatible(source, target):
681
"""Be compatible with GitRepository."""
682
return (isinstance(source, LocalGitRepository) and
683
isinstance(target, LocalGitRepository))
686
class InterRemoteGitLocalGitRepository(InterGitGitRepository):
688
def fetch_objects(self, determine_wants, mapping=None, limit=None, lossy=False):
690
raise LossyPushToSameVCS(self.source, self.target)
691
if limit is not None:
692
raise FetchLimitUnsupported(self)
693
graphwalker = self.target._git.get_graph_walker()
694
pb = ui.ui_factory.nested_progress_bar()
696
if CAPABILITY_THIN_PACK in self.source.controldir._client._fetch_capabilities:
697
# TODO(jelmer): Avoid reading entire file into memory and
698
# only processing it after the whole file has been fetched.
704
self.target._git.object_store.move_in_thin_pack(f)
709
f, commit, abort = self.target._git.object_store.add_pack()
711
refs = self.source.controldir.fetch_pack(
712
determine_wants, graphwalker, f.write,
713
lambda text: report_git_progress(pb, text))
715
return (None, None, refs)
716
except BaseException:
723
def is_compatible(source, target):
724
"""Be compatible with GitRepository."""
725
return (isinstance(source, RemoteGitRepository) and
726
isinstance(target, LocalGitRepository))