202
191
self.random_revid = False
204
def will_record_deletes(self):
205
"""Tell the commit builder that deletes are being notified.
207
This enables the accumulation of an inventory delta; for the resulting
208
commit to be valid, deletes against the basis MUST be recorded via
209
builder.record_delete().
211
raise NotImplementedError(self.will_record_deletes)
213
def record_iter_changes(self, tree, basis_revision_id, iter_changes):
214
"""Record a new tree via iter_changes.
216
:param tree: The tree to obtain text contents from for changed objects.
217
:param basis_revision_id: The revision id of the tree the iter_changes
218
has been generated against. Currently assumed to be the same
219
as self.parents[0] - if it is not, errors may occur.
220
:param iter_changes: An iter_changes iterator with the changes to apply
221
to basis_revision_id. The iterator must not include any items with
222
a current kind of None - missing items must be either filtered out
223
or errored-on beefore record_iter_changes sees the item.
224
:return: A generator of (file_id, relpath, fs_hash) tuples for use with
227
raise NotImplementedError(self.record_iter_changes)
230
class RepositoryWriteLockResult(LogicalLockResult):
231
"""The result of write locking a repository.
233
:ivar repository_token: The token obtained from the underlying lock, or
235
:ivar unlock: A callable which will unlock the lock.
238
def __init__(self, unlock, repository_token):
239
LogicalLockResult.__init__(self, unlock)
240
self.repository_token = repository_token
243
return "RepositoryWriteLockResult(%s, %s)" % (self.repository_token,
193
def _heads(self, file_id, revision_ids):
194
"""Calculate the graph heads for revision_ids in the graph of file_id.
196
This can use either a per-file graph or a global revision graph as we
197
have an identity relationship between the two graphs.
199
return self.__heads(revision_ids)
201
def _check_root(self, ie, parent_invs, tree):
202
"""Helper for record_entry_contents.
204
:param ie: An entry being added.
205
:param parent_invs: The inventories of the parent revisions of the
207
:param tree: The tree that is being committed.
209
# In this revision format, root entries have no knit or weave When
210
# serializing out to disk and back in root.revision is always
212
ie.revision = self._new_revision_id
214
def _get_delta(self, ie, basis_inv, path):
215
"""Get a delta against the basis inventory for ie."""
216
if ie.file_id not in basis_inv:
218
return (None, path, ie.file_id, ie)
219
elif ie != basis_inv[ie.file_id]:
221
# TODO: avoid tis id2path call.
222
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
227
def record_entry_contents(self, ie, parent_invs, path, tree,
229
"""Record the content of ie from tree into the commit if needed.
231
Side effect: sets ie.revision when unchanged
233
:param ie: An inventory entry present in the commit.
234
:param parent_invs: The inventories of the parent revisions of the
236
:param path: The path the entry is at in the tree.
237
:param tree: The tree which contains this entry and should be used to
239
:param content_summary: Summary data from the tree about the paths
240
content - stat, length, exec, sha/link target. This is only
241
accessed when the entry has a revision of None - that is when it is
242
a candidate to commit.
243
:return: A tuple (change_delta, version_recorded). change_delta is
244
an inventory_delta change for this entry against the basis tree of
245
the commit, or None if no change occured against the basis tree.
246
version_recorded is True if a new version of the entry has been
247
recorded. For instance, committing a merge where a file was only
248
changed on the other side will return (delta, False).
250
if self.new_inventory.root is None:
251
if ie.parent_id is not None:
252
raise errors.RootMissing()
253
self._check_root(ie, parent_invs, tree)
254
if ie.revision is None:
255
kind = content_summary[0]
257
# ie is carried over from a prior commit
259
# XXX: repository specific check for nested tree support goes here - if
260
# the repo doesn't want nested trees we skip it ?
261
if (kind == 'tree-reference' and
262
not self.repository._format.supports_tree_reference):
263
# mismatch between commit builder logic and repository:
264
# this needs the entry creation pushed down into the builder.
265
raise NotImplementedError('Missing repository subtree support.')
266
self.new_inventory.add(ie)
268
# TODO: slow, take it out of the inner loop.
270
basis_inv = parent_invs[0]
272
basis_inv = Inventory(root_id=None)
274
# ie.revision is always None if the InventoryEntry is considered
275
# for committing. We may record the previous parents revision if the
276
# content is actually unchanged against a sole head.
277
if ie.revision is not None:
278
if not self._versioned_root and path == '':
279
# repositories that do not version the root set the root's
280
# revision to the new commit even when no change occurs, and
281
# this masks when a change may have occurred against the basis,
282
# so calculate if one happened.
283
if ie.file_id in basis_inv:
284
delta = (basis_inv.id2path(ie.file_id), path,
288
delta = (None, path, ie.file_id, ie)
291
# we don't need to commit this, because the caller already
292
# determined that an existing revision of this file is
294
return None, (ie.revision == self._new_revision_id)
295
# XXX: Friction: parent_candidates should return a list not a dict
296
# so that we don't have to walk the inventories again.
297
parent_candiate_entries = ie.parent_candidates(parent_invs)
298
head_set = self._heads(ie.file_id, parent_candiate_entries.keys())
300
for inv in parent_invs:
301
if ie.file_id in inv:
302
old_rev = inv[ie.file_id].revision
303
if old_rev in head_set:
304
heads.append(inv[ie.file_id].revision)
305
head_set.remove(inv[ie.file_id].revision)
308
# now we check to see if we need to write a new record to the
310
# We write a new entry unless there is one head to the ancestors, and
311
# the kind-derived content is unchanged.
313
# Cheapest check first: no ancestors, or more the one head in the
314
# ancestors, we write a new node.
318
# There is a single head, look it up for comparison
319
parent_entry = parent_candiate_entries[heads[0]]
320
# if the non-content specific data has changed, we'll be writing a
322
if (parent_entry.parent_id != ie.parent_id or
323
parent_entry.name != ie.name):
325
# now we need to do content specific checks:
327
# if the kind changed the content obviously has
328
if kind != parent_entry.kind:
331
assert content_summary[2] is not None, \
332
"Files must not have executable = None"
334
if (# if the file length changed we have to store:
335
parent_entry.text_size != content_summary[1] or
336
# if the exec bit has changed we have to store:
337
parent_entry.executable != content_summary[2]):
339
elif parent_entry.text_sha1 == content_summary[3]:
340
# all meta and content is unchanged (using a hash cache
341
# hit to check the sha)
342
ie.revision = parent_entry.revision
343
ie.text_size = parent_entry.text_size
344
ie.text_sha1 = parent_entry.text_sha1
345
ie.executable = parent_entry.executable
346
return self._get_delta(ie, basis_inv, path), False
348
# Either there is only a hash change(no hash cache entry,
349
# or same size content change), or there is no change on
351
# Provide the parent's hash to the store layer, so that the
352
# content is unchanged we will not store a new node.
353
nostore_sha = parent_entry.text_sha1
355
# We want to record a new node regardless of the presence or
356
# absence of a content change in the file.
358
ie.executable = content_summary[2]
359
lines = tree.get_file(ie.file_id, path).readlines()
361
ie.text_sha1, ie.text_size = self._add_text_to_weave(
362
ie.file_id, lines, heads, nostore_sha)
363
except errors.ExistingContent:
364
# Turns out that the file content was unchanged, and we were
365
# only going to store a new node if it was changed. Carry over
367
ie.revision = parent_entry.revision
368
ie.text_size = parent_entry.text_size
369
ie.text_sha1 = parent_entry.text_sha1
370
ie.executable = parent_entry.executable
371
return self._get_delta(ie, basis_inv, path), False
372
elif kind == 'directory':
374
# all data is meta here, nothing specific to directory, so
376
ie.revision = parent_entry.revision
377
return self._get_delta(ie, basis_inv, path), False
379
self._add_text_to_weave(ie.file_id, lines, heads, None)
380
elif kind == 'symlink':
381
current_link_target = content_summary[3]
383
# symlink target is not generic metadata, check if it has
385
if current_link_target != parent_entry.symlink_target:
388
# unchanged, carry over.
389
ie.revision = parent_entry.revision
390
ie.symlink_target = parent_entry.symlink_target
391
return self._get_delta(ie, basis_inv, path), False
392
ie.symlink_target = current_link_target
394
self._add_text_to_weave(ie.file_id, lines, heads, None)
395
elif kind == 'tree-reference':
397
if content_summary[3] != parent_entry.reference_revision:
400
# unchanged, carry over.
401
ie.reference_revision = parent_entry.reference_revision
402
ie.revision = parent_entry.revision
403
return self._get_delta(ie, basis_inv, path), False
404
ie.reference_revision = content_summary[3]
406
self._add_text_to_weave(ie.file_id, lines, heads, None)
408
raise NotImplementedError('unknown kind')
409
ie.revision = self._new_revision_id
410
return self._get_delta(ie, basis_inv, path), True
412
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
413
versionedfile = self.repository.weave_store.get_weave_or_empty(
414
file_id, self.repository.get_transaction())
415
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
416
# than add_lines, and allows committing when a parent is ghosted for
418
# Note: as we read the content directly from the tree, we know its not
419
# been turned into unicode or badly split - but a broken tree
420
# implementation could give us bad output from readlines() so this is
421
# not a guarantee of safety. What would be better is always checking
422
# the content during test suite execution. RBC 20070912
423
return versionedfile.add_lines_with_ghosts(
424
self._new_revision_id, parents, new_lines,
425
nostore_sha=nostore_sha, random_id=self.random_revid,
426
check_content=False)[0:2]
429
class RootCommitBuilder(CommitBuilder):
430
"""This commitbuilder actually records the root id"""
432
# the root entry gets versioned properly by this builder.
433
_versioned_root = True
435
def _check_root(self, ie, parent_invs, tree):
436
"""Helper for record_entry_contents.
438
:param ie: An entry being added.
439
:param parent_invs: The inventories of the parent revisions of the
441
:param tree: The tree that is being committed.
247
445
######################################################################
251
class Repository(_RelockDebugMixin, controldir.ControlComponent):
448
class Repository(object):
252
449
"""Repository holding history for one or more branches.
254
451
The repository holds and retrieves historical information including
255
452
revisions and file history. It's normally accessed only by the Branch,
256
453
which views a particular line of development through that history.
258
See VersionedFileRepository in bzrlib.vf_repository for the
259
base class for most Bazaar repositories.
455
The Repository builds on top of Stores and a Transport, which respectively
456
describe the disk data format and the way of accessing the (possibly
262
def abort_write_group(self, suppress_errors=False):
460
# What class to use for a CommitBuilder. Often its simpler to change this
461
# in a Repository class subclass rather than to override
462
# get_commit_builder.
463
_commit_builder_class = CommitBuilder
464
# The search regex used by xml based repositories to determine what things
465
# where changed in a single commit.
466
_file_ids_altered_regex = lazy_regex.lazy_compile(
467
r'file_id="(?P<file_id>[^"]+)"'
468
r'.* revision="(?P<revision_id>[^"]+)"'
471
def abort_write_group(self):
263
472
"""Commit the contents accrued within the current write group.
265
:param suppress_errors: if true, abort_write_group will catch and log
266
unexpected errors that happen during the abort, rather than
267
allowing them to propagate. Defaults to False.
269
474
:seealso: start_write_group.
271
476
if self._write_group is not self.get_transaction():
272
477
# has an unlock or relock occured ?
275
'(suppressed) mismatched lock context and write group. %r, %r',
276
self._write_group, self.get_transaction())
278
raise errors.BzrError(
279
'mismatched lock context and write group. %r, %r' %
280
(self._write_group, self.get_transaction()))
282
self._abort_write_group()
283
except Exception, exc:
284
self._write_group = None
285
if not suppress_errors:
287
mutter('abort_write_group failed')
288
log_exception_quietly()
289
note(gettext('bzr: ERROR (ignored): %s'), exc)
478
raise errors.BzrError('mismatched lock context and write group.')
479
self._abort_write_group()
290
480
self._write_group = None
292
482
def _abort_write_group(self):
293
483
"""Template method for per-repository write group cleanup.
295
This is called during abort before the write group is considered to be
485
This is called during abort before the write group is considered to be
296
486
finished and should cleanup any internal state accrued during the write
297
487
group. There is no requirement that data handed to the repository be
298
488
*not* made available - this is not a rollback - but neither should any
546
771
:param using: If True, list only branches using this repository.
548
773
if using and not self.is_shared():
549
return self.bzrdir.list_branches()
775
return [self.bzrdir.open_branch()]
776
except errors.NotBranchError:
550
778
class Evaluator(object):
552
780
def __init__(self):
553
781
self.first_call = True
555
def __call__(self, controldir):
556
# On the first call, the parameter is always the controldir
783
def __call__(self, bzrdir):
784
# On the first call, the parameter is always the bzrdir
557
785
# containing the current repo.
558
786
if not self.first_call:
560
repository = controldir.open_repository()
788
repository = bzrdir.open_repository()
561
789
except errors.NoRepositoryPresent:
564
return False, ([], repository)
792
return False, (None, repository)
565
793
self.first_call = False
566
value = (controldir.list_branches(), None)
795
value = (bzrdir.open_branch(), None)
796
except errors.NotBranchError:
567
798
return True, value
570
for branches, repository in controldir.ControlDir.find_bzrdirs(
571
self.user_transport, evaluate=Evaluator()):
572
if branches is not None:
801
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
802
self.bzrdir.root_transport, evaluate=Evaluator()):
803
if branch is not None:
804
branches.append(branch)
574
805
if not using and repository is not None:
575
ret.extend(repository.find_branches())
806
branches.extend(repository.find_branches())
809
def get_data_stream(self, revision_ids):
810
raise NotImplementedError(self.get_data_stream)
812
def get_data_stream_for_search(self, search_result):
813
"""Get a data stream that can be inserted to a repository.
815
:param search_result: A bzrlib.graph.SearchResult selecting the
817
:return: A data stream that can be inserted into a repository using
820
raise NotImplementedError(self.get_data_stream_for_search)
822
def insert_data_stream(self, stream):
823
"""XXX What does this really do?
825
Is it a substitute for fetch?
826
Should it manage its own write group ?
828
for item_key, bytes in stream:
829
if item_key[0] == 'file':
830
(file_id,) = item_key[1:]
831
knit = self.weave_store.get_weave_or_empty(
832
file_id, self.get_transaction())
833
elif item_key == ('inventory',):
834
knit = self.get_inventory_weave()
835
elif item_key == ('revisions',):
836
knit = self._revision_store.get_revision_file(
837
self.get_transaction())
838
elif item_key == ('signatures',):
839
knit = self._revision_store.get_signature_file(
840
self.get_transaction())
842
raise errors.RepositoryDataStreamError(
843
"Unrecognised data stream key '%s'" % (item_key,))
844
decoded_list = bencode.bdecode(bytes)
845
format = decoded_list.pop(0)
848
for version, options, parents, some_bytes in decoded_list:
849
data_list.append((version, options, len(some_bytes), parents))
850
knit_bytes += some_bytes
851
buffer = StringIO(knit_bytes)
852
def reader_func(count):
856
return buffer.read(count)
857
knit.insert_data_stream(
858
(format, data_list, reader_func))
579
def search_missing_revision_ids(self, other,
580
revision_id=symbol_versioning.DEPRECATED_PARAMETER,
581
find_ghosts=True, revision_ids=None, if_present_ids=None,
861
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
583
862
"""Return the revision ids that other has that this does not.
585
864
These are returned in topological order.
587
866
revision_id: only return revision ids included by revision_id.
589
if symbol_versioning.deprecated_passed(revision_id):
590
symbol_versioning.warn(
591
'search_missing_revision_ids(revision_id=...) was '
592
'deprecated in 2.4. Use revision_ids=[...] instead.',
593
DeprecationWarning, stacklevel=3)
594
if revision_ids is not None:
595
raise AssertionError(
596
'revision_ids is mutually exclusive with revision_id')
597
if revision_id is not None:
598
revision_ids = [revision_id]
599
868
return InterRepository.get(other, self).search_missing_revision_ids(
600
find_ghosts=find_ghosts, revision_ids=revision_ids,
601
if_present_ids=if_present_ids, limit=limit)
869
revision_id, find_ghosts)
871
@deprecated_method(symbol_versioning.one_two)
873
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
874
"""Return the revision ids that other has that this does not.
876
These are returned in topological order.
878
revision_id: only return revision ids included by revision_id.
880
keys = self.search_missing_revision_ids(
881
other, revision_id, find_ghosts).get_keys()
884
parents = other.get_graph().get_parent_map(keys)
887
return tsort.topo_sort(parents)
828
1061
raise NotImplementedError(self.has_revisions)
1063
return self._revision_store.has_revision_id(revision_id,
1064
self.get_transaction())
830
1066
@needs_read_lock
831
1067
def get_revision(self, revision_id):
832
1068
"""Return the Revision object for a named revision."""
833
1069
return self.get_revisions([revision_id])[0]
835
1072
def get_revision_reconcile(self, revision_id):
836
1073
"""'reconcile' helper routine that allows access to a revision always.
838
1075
This variant of get_revision does not cross check the weave graph
839
1076
against the revision one as get_revision does: but it should only
840
1077
be used by reconcile, or reconcile-alike commands that are correcting
841
1078
or testing the revision graph.
843
raise NotImplementedError(self.get_revision_reconcile)
1080
return self._get_revisions([revision_id])[0]
845
1083
def get_revisions(self, revision_ids):
846
"""Get many revisions at once.
848
Repositories that need to check data on every revision read should
849
subclass this method.
851
raise NotImplementedError(self.get_revisions)
853
def get_deltas_for_revisions(self, revisions, specific_fileids=None):
1084
"""Get many revisions at once."""
1085
return self._get_revisions(revision_ids)
1088
def _get_revisions(self, revision_ids):
1089
"""Core work logic to get many revisions without sanity checks."""
1090
for rev_id in revision_ids:
1091
if not rev_id or not isinstance(rev_id, basestring):
1092
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1093
revs = self._revision_store.get_revisions(revision_ids,
1094
self.get_transaction())
1096
assert not isinstance(rev.revision_id, unicode)
1097
for parent_id in rev.parent_ids:
1098
assert not isinstance(parent_id, unicode)
1102
def get_revision_xml(self, revision_id):
1103
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1104
# would have already do it.
1105
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1106
rev = self.get_revision(revision_id)
1107
rev_tmp = StringIO()
1108
# the current serializer..
1109
self._revision_store._serializer.write_revision(rev, rev_tmp)
1111
return rev_tmp.getvalue()
1114
def get_deltas_for_revisions(self, revisions):
854
1115
"""Produce a generator of revision deltas.
856
1117
Note that the input is a sequence of REVISIONS, not revision_ids.
857
1118
Trees will be held in memory until the generator exits.
858
1119
Each delta is relative to the revision's lefthand predecessor.
860
:param specific_fileids: if not None, the result is filtered
861
so that only those file-ids, their parents and their
862
children are included.
864
# Get the revision-ids of interest
865
1121
required_trees = set()
866
1122
for revision in revisions:
867
1123
required_trees.add(revision.revision_id)
868
1124
required_trees.update(revision.parent_ids[:1])
870
# Get the matching filtered trees. Note that it's more
871
# efficient to pass filtered trees to changes_from() rather
872
# than doing the filtering afterwards. changes_from() could
873
# arguably do the filtering itself but it's path-based, not
874
# file-id based, so filtering before or afterwards is
876
if specific_fileids is None:
877
trees = dict((t.get_revision_id(), t) for
878
t in self.revision_trees(required_trees))
880
trees = dict((t.get_revision_id(), t) for
881
t in self._filtered_revision_trees(required_trees,
884
# Calculate the deltas
1125
trees = dict((t.get_revision_id(), t) for
1126
t in self.revision_trees(required_trees))
885
1127
for revision in revisions:
886
1128
if not revision.parent_ids:
887
old_tree = self.revision_tree(_mod_revision.NULL_REVISION)
1129
old_tree = self.revision_tree(None)
889
1131
old_tree = trees[revision.parent_ids[0]]
890
1132
yield trees[revision.revision_id].changes_from(old_tree)
892
1134
@needs_read_lock
893
def get_revision_delta(self, revision_id, specific_fileids=None):
1135
def get_revision_delta(self, revision_id):
894
1136
"""Return the delta for one revision.
896
1138
The delta is relative to the left-hand predecessor of the
899
:param specific_fileids: if not None, the result is filtered
900
so that only those file-ids, their parents and their
901
children are included.
903
1141
r = self.get_revision(revision_id)
904
return list(self.get_deltas_for_revisions([r],
905
specific_fileids=specific_fileids))[0]
1142
return list(self.get_deltas_for_revisions([r]))[0]
907
1144
@needs_write_lock
908
1145
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
909
1146
signature = gpg_strategy.sign(plaintext)
910
1147
self.add_signature_text(revision_id, signature)
912
1150
def add_signature_text(self, revision_id, signature):
913
"""Store a signature text for a revision.
915
:param revision_id: Revision id of the revision
916
:param signature: Signature text.
918
raise NotImplementedError(self.add_signature_text)
920
def _find_parent_ids_of_revisions(self, revision_ids):
921
"""Find all parent ids that are mentioned in the revision graph.
923
:return: set of revisions that are parents of revision_ids which are
924
not part of revision_ids themselves
926
parent_map = self.get_parent_map(revision_ids)
928
map(parent_ids.update, parent_map.itervalues())
929
parent_ids.difference_update(revision_ids)
930
parent_ids.discard(_mod_revision.NULL_REVISION)
1151
self._revision_store.add_revision_signature_text(revision_id,
1153
self.get_transaction())
1155
def find_text_key_references(self):
1156
"""Find the text key references within the repository.
1158
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1159
revision_ids. Each altered file-ids has the exact revision_ids that
1160
altered it listed explicitly.
1161
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1162
to whether they were referred to by the inventory of the
1163
revision_id that they contain. The inventory texts from all present
1164
revision ids are assessed to generate this report.
1166
revision_ids = self.all_revision_ids()
1167
w = self.get_inventory_weave()
1168
pb = ui.ui_factory.nested_progress_bar()
1170
return self._find_text_key_references_from_xml_inventory_lines(
1171
w.iter_lines_added_or_present_in_versions(revision_ids, pb=pb))
1175
def _find_text_key_references_from_xml_inventory_lines(self,
1177
"""Core routine for extracting references to texts from inventories.
1179
This performs the translation of xml lines to revision ids.
1181
:param line_iterator: An iterator of lines, origin_version_id
1182
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1183
to whether they were referred to by the inventory of the
1184
revision_id that they contain. Note that if that revision_id was
1185
not part of the line_iterator's output then False will be given -
1186
even though it may actually refer to that key.
1188
if not self._serializer.support_altered_by_hack:
1189
raise AssertionError(
1190
"_find_text_key_references_from_xml_inventory_lines only "
1191
"supported for branches which store inventory as unnested xml"
1192
", not on %r" % self)
1195
# this code needs to read every new line in every inventory for the
1196
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1197
# not present in one of those inventories is unnecessary but not
1198
# harmful because we are filtering by the revision id marker in the
1199
# inventory lines : we only select file ids altered in one of those
1200
# revisions. We don't need to see all lines in the inventory because
1201
# only those added in an inventory in rev X can contain a revision=X
1203
unescape_revid_cache = {}
1204
unescape_fileid_cache = {}
1206
# jam 20061218 In a big fetch, this handles hundreds of thousands
1207
# of lines, so it has had a lot of inlining and optimizing done.
1208
# Sorry that it is a little bit messy.
1209
# Move several functions to be local variables, since this is a long
1211
search = self._file_ids_altered_regex.search
1212
unescape = _unescape_xml
1213
setdefault = result.setdefault
1214
for line, version_id in line_iterator:
1215
match = search(line)
1218
# One call to match.group() returning multiple items is quite a
1219
# bit faster than 2 calls to match.group() each returning 1
1220
file_id, revision_id = match.group('file_id', 'revision_id')
1222
# Inlining the cache lookups helps a lot when you make 170,000
1223
# lines and 350k ids, versus 8.4 unique ids.
1224
# Using a cache helps in 2 ways:
1225
# 1) Avoids unnecessary decoding calls
1226
# 2) Re-uses cached strings, which helps in future set and
1228
# (2) is enough that removing encoding entirely along with
1229
# the cache (so we are using plain strings) results in no
1230
# performance improvement.
1232
revision_id = unescape_revid_cache[revision_id]
1234
unescaped = unescape(revision_id)
1235
unescape_revid_cache[revision_id] = unescaped
1236
revision_id = unescaped
1238
# Note that unconditionally unescaping means that we deserialise
1239
# every fileid, which for general 'pull' is not great, but we don't
1240
# really want to have some many fulltexts that this matters anyway.
1243
file_id = unescape_fileid_cache[file_id]
1245
unescaped = unescape(file_id)
1246
unescape_fileid_cache[file_id] = unescaped
1249
key = (file_id, revision_id)
1250
setdefault(key, False)
1251
if revision_id == version_id:
1255
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1257
"""Helper routine for fileids_altered_by_revision_ids.
1259
This performs the translation of xml lines to revision ids.
1261
:param line_iterator: An iterator of lines, origin_version_id
1262
:param revision_ids: The revision ids to filter for. This should be a
1263
set or other type which supports efficient __contains__ lookups, as
1264
the revision id from each parsed line will be looked up in the
1265
revision_ids filter.
1266
:return: a dictionary mapping altered file-ids to an iterable of
1267
revision_ids. Each altered file-ids has the exact revision_ids that
1268
altered it listed explicitly.
1271
setdefault = result.setdefault
1272
for file_id, revision_id in \
1273
self._find_text_key_references_from_xml_inventory_lines(
1274
line_iterator).iterkeys():
1275
# once data is all ensured-consistent; then this is
1276
# if revision_id == version_id
1277
if revision_id in revision_ids:
1278
setdefault(file_id, set()).add(revision_id)
1281
def fileids_altered_by_revision_ids(self, revision_ids):
1282
"""Find the file ids and versions affected by revisions.
1284
:param revisions: an iterable containing revision ids.
1285
:return: a dictionary mapping altered file-ids to an iterable of
1286
revision_ids. Each altered file-ids has the exact revision_ids that
1287
altered it listed explicitly.
1289
selected_revision_ids = set(revision_ids)
1290
w = self.get_inventory_weave()
1291
pb = ui.ui_factory.nested_progress_bar()
1293
return self._find_file_ids_from_xml_inventory_lines(
1294
w.iter_lines_added_or_present_in_versions(
1295
selected_revision_ids, pb=pb),
1296
selected_revision_ids)
933
1300
def iter_files_bytes(self, desired_files):
934
1301
"""Iterate through file versions.
941
1308
uniquely identify the file version in the caller's context. (Examples:
942
1309
an index number or a TreeTransform trans_id.)
1311
bytes_iterator is an iterable of bytestrings for the file. The
1312
kind of iterable and length of the bytestrings are unspecified, but for
1313
this implementation, it is a list of lines produced by
1314
VersionedFile.get_lines().
944
1316
:param desired_files: a list of (file_id, revision_id, identifier)
947
raise NotImplementedError(self.iter_files_bytes)
949
def get_rev_id_for_revno(self, revno, known_pair):
950
"""Return the revision id of a revno, given a later (revno, revid)
951
pair in the same history.
953
:return: if found (True, revid). If the available history ran out
954
before reaching the revno, then this returns
955
(False, (closest_revno, closest_revid)).
957
known_revno, known_revid = known_pair
958
partial_history = [known_revid]
959
distance_from_known = known_revno - revno
960
if distance_from_known < 0:
962
'requested revno (%d) is later than given known revno (%d)'
963
% (revno, known_revno))
966
self, partial_history, stop_index=distance_from_known)
967
except errors.RevisionNotPresent, err:
968
if err.revision_id == known_revid:
969
# The start revision (known_revid) wasn't found.
971
# This is a stacked repository with no fallbacks, or a there's a
972
# left-hand ghost. Either way, even though the revision named in
973
# the error isn't in this repo, we know it's the next step in this
975
partial_history.append(err.revision_id)
976
if len(partial_history) <= distance_from_known:
977
# Didn't find enough history to get a revid for the revno.
978
earliest_revno = known_revno - len(partial_history) + 1
979
return (False, (earliest_revno, partial_history[-1]))
980
if len(partial_history) - 1 > distance_from_known:
981
raise AssertionError('_iter_for_revno returned too much history')
982
return (True, partial_history[-1])
1319
transaction = self.get_transaction()
1320
for file_id, revision_id, callable_data in desired_files:
1322
weave = self.weave_store.get_weave(file_id, transaction)
1323
except errors.NoSuchFile:
1324
raise errors.NoSuchIdInRepository(self, file_id)
1325
yield callable_data, weave.get_lines(revision_id)
1327
def _generate_text_key_index(self, text_key_references=None,
1329
"""Generate a new text key index for the repository.
1331
This is an expensive function that will take considerable time to run.
1333
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1334
list of parents, also text keys. When a given key has no parents,
1335
the parents list will be [NULL_REVISION].
1337
# All revisions, to find inventory parents.
1338
if ancestors is None:
1339
graph = self.get_graph()
1340
ancestors = graph.get_parent_map(self.all_revision_ids())
1341
if text_key_references is None:
1342
text_key_references = self.find_text_key_references()
1343
pb = ui.ui_factory.nested_progress_bar()
1345
return self._do_generate_text_key_index(ancestors,
1346
text_key_references, pb)
1350
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1351
"""Helper for _generate_text_key_index to avoid deep nesting."""
1352
revision_order = tsort.topo_sort(ancestors)
1353
invalid_keys = set()
1355
for revision_id in revision_order:
1356
revision_keys[revision_id] = set()
1357
text_count = len(text_key_references)
1358
# a cache of the text keys to allow reuse; costs a dict of all the
1359
# keys, but saves a 2-tuple for every child of a given key.
1361
for text_key, valid in text_key_references.iteritems():
1363
invalid_keys.add(text_key)
1365
revision_keys[text_key[1]].add(text_key)
1366
text_key_cache[text_key] = text_key
1367
del text_key_references
1369
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1370
NULL_REVISION = _mod_revision.NULL_REVISION
1371
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1372
# too small for large or very branchy trees. However, for 55K path
1373
# trees, it would be easy to use too much memory trivially. Ideally we
1374
# could gauge this by looking at available real memory etc, but this is
1375
# always a tricky proposition.
1376
inventory_cache = lru_cache.LRUCache(10)
1377
batch_size = 10 # should be ~150MB on a 55K path tree
1378
batch_count = len(revision_order) / batch_size + 1
1380
pb.update("Calculating text parents.", processed_texts, text_count)
1381
for offset in xrange(batch_count):
1382
to_query = revision_order[offset * batch_size:(offset + 1) *
1386
for rev_tree in self.revision_trees(to_query):
1387
revision_id = rev_tree.get_revision_id()
1388
parent_ids = ancestors[revision_id]
1389
for text_key in revision_keys[revision_id]:
1390
pb.update("Calculating text parents.", processed_texts)
1391
processed_texts += 1
1392
candidate_parents = []
1393
for parent_id in parent_ids:
1394
parent_text_key = (text_key[0], parent_id)
1396
check_parent = parent_text_key not in \
1397
revision_keys[parent_id]
1399
# the parent parent_id is a ghost:
1400
check_parent = False
1401
# truncate the derived graph against this ghost.
1402
parent_text_key = None
1404
# look at the parent commit details inventories to
1405
# determine possible candidates in the per file graph.
1408
inv = inventory_cache[parent_id]
1410
inv = self.revision_tree(parent_id).inventory
1411
inventory_cache[parent_id] = inv
1412
parent_entry = inv._byid.get(text_key[0], None)
1413
if parent_entry is not None:
1415
text_key[0], parent_entry.revision)
1417
parent_text_key = None
1418
if parent_text_key is not None:
1419
candidate_parents.append(
1420
text_key_cache[parent_text_key])
1421
parent_heads = text_graph.heads(candidate_parents)
1422
new_parents = list(parent_heads)
1423
new_parents.sort(key=lambda x:candidate_parents.index(x))
1424
if new_parents == []:
1425
new_parents = [NULL_REVISION]
1426
text_index[text_key] = new_parents
1428
for text_key in invalid_keys:
1429
text_index[text_key] = [NULL_REVISION]
1432
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1433
"""Get an iterable listing the keys of all the data introduced by a set
1436
The keys will be ordered so that the corresponding items can be safely
1437
fetched and inserted in that order.
1439
:returns: An iterable producing tuples of (knit-kind, file-id,
1440
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1441
'revisions'. file-id is None unless knit-kind is 'file'.
1443
# XXX: it's a bit weird to control the inventory weave caching in this
1444
# generator. Ideally the caching would be done in fetch.py I think. Or
1445
# maybe this generator should explicitly have the contract that it
1446
# should not be iterated until the previously yielded item has been
1449
inv_w = self.get_inventory_weave()
1451
# file ids that changed
1452
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1454
num_file_ids = len(file_ids)
1455
for file_id, altered_versions in file_ids.iteritems():
1456
if _files_pb is not None:
1457
_files_pb.update("fetch texts", count, num_file_ids)
1459
yield ("file", file_id, altered_versions)
1460
# We're done with the files_pb. Note that it finished by the caller,
1461
# just as it was created by the caller.
1465
yield ("inventory", None, revision_ids)
1468
revisions_with_signatures = set()
1469
for rev_id in revision_ids:
1471
self.get_signature_text(rev_id)
1472
except errors.NoSuchRevision:
1476
revisions_with_signatures.add(rev_id)
1478
yield ("signatures", None, revisions_with_signatures)
1481
yield ("revisions", None, revision_ids)
1484
def get_inventory_weave(self):
1485
return self.control_weaves.get_weave('inventory',
1486
self.get_transaction())
1489
def get_inventory(self, revision_id):
1490
"""Get Inventory object by revision id."""
1491
return self.iter_inventories([revision_id]).next()
1493
def iter_inventories(self, revision_ids):
1494
"""Get many inventories by revision_ids.
1496
This will buffer some or all of the texts used in constructing the
1497
inventories in memory, but will only parse a single inventory at a
1500
:return: An iterator of inventories.
1502
assert None not in revision_ids
1503
assert _mod_revision.NULL_REVISION not in revision_ids
1504
return self._iter_inventories(revision_ids)
1506
def _iter_inventories(self, revision_ids):
1507
"""single-document based inventory iteration."""
1508
texts = self.get_inventory_weave().get_texts(revision_ids)
1509
for text, revision_id in zip(texts, revision_ids):
1510
yield self.deserialise_inventory(revision_id, text)
1512
def deserialise_inventory(self, revision_id, xml):
1513
"""Transform the xml into an inventory object.
1515
:param revision_id: The expected revision id of the inventory.
1516
:param xml: A serialised inventory.
1518
result = self._serializer.read_inventory_from_string(xml, revision_id)
1519
if result.revision_id != revision_id:
1520
raise AssertionError('revision id mismatch %s != %s' % (
1521
result.revision_id, revision_id))
1524
def serialise_inventory(self, inv):
1525
return self._serializer.write_inventory_to_string(inv)
1527
def _serialise_inventory_to_lines(self, inv):
1528
return self._serializer.write_inventory_to_lines(inv)
1530
def get_serializer_format(self):
1531
return self._serializer.format_num
1534
def get_inventory_xml(self, revision_id):
1535
"""Get inventory XML as a file object."""
1537
assert isinstance(revision_id, str), type(revision_id)
1538
iw = self.get_inventory_weave()
1539
return iw.get_text(revision_id)
1541
raise errors.HistoryMissing(self, 'inventory', revision_id)
1544
def get_inventory_sha1(self, revision_id):
1545
"""Return the sha1 hash of the inventory entry
1547
return self.get_revision(revision_id).inventory_sha1
1550
@deprecated_method(symbol_versioning.one_four)
1551
def get_revision_graph(self, revision_id=None):
1552
"""Return a dictionary containing the revision graph.
1554
NB: This method should not be used as it accesses the entire graph all
1555
at once, which is much more data than most operations should require.
1557
:param revision_id: The revision_id to get a graph from. If None, then
1558
the entire revision graph is returned. This is a deprecated mode of
1559
operation and will be removed in the future.
1560
:return: a dictionary of revision_id->revision_parents_list.
1562
raise NotImplementedError(self.get_revision_graph)
1565
@deprecated_method(symbol_versioning.one_three)
1566
def get_revision_graph_with_ghosts(self, revision_ids=None):
1567
"""Return a graph of the revisions with ghosts marked as applicable.
1569
:param revision_ids: an iterable of revisions to graph or None for all.
1570
:return: a Graph object with the graph reachable from revision_ids.
1572
if 'evil' in debug.debug_flags:
1574
"get_revision_graph_with_ghosts scales with size of history.")
1575
result = deprecated_graph.Graph()
1576
if not revision_ids:
1577
pending = set(self.all_revision_ids())
1580
pending = set(revision_ids)
1581
# special case NULL_REVISION
1582
if _mod_revision.NULL_REVISION in pending:
1583
pending.remove(_mod_revision.NULL_REVISION)
1584
required = set(pending)
1587
revision_id = pending.pop()
1589
rev = self.get_revision(revision_id)
1590
except errors.NoSuchRevision:
1591
if revision_id in required:
1594
result.add_ghost(revision_id)
1596
for parent_id in rev.parent_ids:
1597
# is this queued or done ?
1598
if (parent_id not in pending and
1599
parent_id not in done):
1601
pending.add(parent_id)
1602
result.add_node(revision_id, rev.parent_ids)
1603
done.add(revision_id)
1606
def iter_reverse_revision_history(self, revision_id):
1607
"""Iterate backwards through revision ids in the lefthand history
1609
:param revision_id: The revision id to start with. All its lefthand
1610
ancestors will be traversed.
1612
graph = self.get_graph()
1613
next_id = revision_id
1615
if next_id in (None, _mod_revision.NULL_REVISION):
1618
# Note: The following line may raise KeyError in the event of
1619
# truncated history. We decided not to have a try:except:raise
1620
# RevisionNotPresent here until we see a use for it, because of the
1621
# cost in an inner loop that is by its very nature O(history).
1622
# Robert Collins 20080326
1623
parents = graph.get_parent_map([next_id])[next_id]
1624
if len(parents) == 0:
1627
next_id = parents[0]
1630
def get_revision_inventory(self, revision_id):
1631
"""Return inventory of a past revision."""
1632
# TODO: Unify this with get_inventory()
1633
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1634
# must be the same as its revision, so this is trivial.
1635
if revision_id is None:
1636
# This does not make sense: if there is no revision,
1637
# then it is the current tree inventory surely ?!
1638
# and thus get_root_id() is something that looks at the last
1639
# commit on the branch, and the get_root_id is an inventory check.
1640
raise NotImplementedError
1641
# return Inventory(self.get_root_id())
1643
return self.get_inventory(revision_id)
984
1646
def is_shared(self):
985
1647
"""Return True if this repository is flagged as a shared repository."""
986
1648
raise NotImplementedError(self.is_shared)
1000
1662
for repositories to maintain loaded indices across multiple locks
1001
1663
by checking inside their implementation of this method to see
1002
1664
whether their indices are still valid. This depends of course on
1003
the disk format being validatable in this manner. This method is
1004
also called by the refresh_data() public interface to cause a refresh
1005
to occur while in a write lock so that data inserted by a smart server
1006
push operation is visible on the client's instance of the physical
1665
the disk format being validatable in this manner.
1010
1668
@needs_read_lock
1011
1669
def revision_tree(self, revision_id):
1012
1670
"""Return Tree for a revision on this branch.
1014
`revision_id` may be NULL_REVISION for the empty tree revision.
1672
`revision_id` may be None for the empty tree revision.
1016
raise NotImplementedError(self.revision_tree)
1674
# TODO: refactor this to use an existing revision object
1675
# so we don't need to read it in twice.
1676
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1677
return RevisionTree(self, Inventory(root_id=None),
1678
_mod_revision.NULL_REVISION)
1680
inv = self.get_revision_inventory(revision_id)
1681
return RevisionTree(self, inv, revision_id)
1018
1684
def revision_trees(self, revision_ids):
1019
"""Return Trees for revisions in this repository.
1021
:param revision_ids: a sequence of revision-ids;
1022
a revision-id may not be None or 'null:'
1685
"""Return Tree for a revision on this branch.
1687
`revision_id` may not be None or 'null:'"""
1688
inventories = self.iter_inventories(revision_ids)
1689
for inv in inventories:
1690
yield RevisionTree(self, inv, inv.revision_id)
1693
def get_ancestry(self, revision_id, topo_sorted=True):
1694
"""Return a list of revision-ids integrated by a revision.
1696
The first element of the list is always None, indicating the origin
1697
revision. This might change when we have history horizons, or
1698
perhaps we should have a new API.
1700
This is topologically sorted.
1024
raise NotImplementedError(self.revision_trees)
1702
if _mod_revision.is_null(revision_id):
1704
if not self.has_revision(revision_id):
1705
raise errors.NoSuchRevision(self, revision_id)
1706
w = self.get_inventory_weave()
1707
candidates = w.get_ancestry(revision_id, topo_sorted)
1708
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1026
def pack(self, hint=None, clean_obsolete_packs=False):
1027
1711
"""Compress the data within the repository.
1029
1713
This operation only makes sense for some repository types. For other
1030
1714
types it should be a no-op that just returns.
1032
1716
This stub method does not require a lock, but subclasses should use
1033
@needs_write_lock as this is a long running call it's reasonable to
1717
@needs_write_lock as this is a long running call its reasonable to
1034
1718
implicitly lock for the user.
1036
:param hint: If not supplied, the whole repository is packed.
1037
If supplied, the repository may use the hint parameter as a
1038
hint for the parts of the repository to pack. A hint can be
1039
obtained from the result of commit_write_group(). Out of
1040
date hints are simply ignored, because concurrent operations
1041
can obsolete them rapidly.
1043
:param clean_obsolete_packs: Clean obsolete packs immediately after
1722
def print_file(self, file, revision_id):
1723
"""Print `file` to stdout.
1725
FIXME RBC 20060125 as John Meinel points out this is a bad api
1726
- it writes to stdout, it assumes that that is valid etc. Fix
1727
by creating a new more flexible convenience function.
1729
tree = self.revision_tree(revision_id)
1730
# use inventory as it was in that revision
1731
file_id = tree.inventory.path2id(file)
1733
# TODO: jam 20060427 Write a test for this code path
1734
# it had a bug in it, and was raising the wrong
1736
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1737
tree.print_file(file_id)
1047
1739
def get_transaction(self):
1048
1740
return self.control_files.get_transaction()
1050
def get_parent_map(self, revision_ids):
1051
"""See graph.StackedParentsProvider.get_parent_map"""
1052
raise NotImplementedError(self.get_parent_map)
1054
def _get_parent_map_no_fallbacks(self, revision_ids):
1055
"""Same as Repository.get_parent_map except doesn't query fallbacks."""
1056
# revisions index works in keys; this just works in revisions
1057
# therefore wrap and unwrap
1060
for revision_id in revision_ids:
1742
def revision_parents(self, revision_id):
1743
return self.get_inventory_weave().parent_names(revision_id)
1745
@deprecated_method(symbol_versioning.one_one)
1746
def get_parents(self, revision_ids):
1747
"""See StackedParentsProvider.get_parents"""
1748
parent_map = self.get_parent_map(revision_ids)
1749
return [parent_map.get(r, None) for r in revision_ids]
1751
def get_parent_map(self, keys):
1752
"""See graph._StackedParentsProvider.get_parent_map"""
1754
for revision_id in keys:
1061
1755
if revision_id == _mod_revision.NULL_REVISION:
1062
result[revision_id] = ()
1063
elif revision_id is None:
1064
raise ValueError('get_parent_map(None) is not valid')
1066
query_keys.append((revision_id ,))
1067
vf = self.revisions.without_fallbacks()
1068
for ((revision_id,), parent_keys) in \
1069
vf.get_parent_map(query_keys).iteritems():
1071
result[revision_id] = tuple([parent_revid
1072
for (parent_revid,) in parent_keys])
1074
result[revision_id] = (_mod_revision.NULL_REVISION,)
1756
parent_map[revision_id] = ()
1759
parent_id_list = self.get_revision(revision_id).parent_ids
1760
except errors.NoSuchRevision:
1763
if len(parent_id_list) == 0:
1764
parent_ids = (_mod_revision.NULL_REVISION,)
1766
parent_ids = tuple(parent_id_list)
1767
parent_map[revision_id] = parent_ids
1077
1770
def _make_parents_provider(self):
1078
if not self._format.supports_external_lookups:
1080
return graph.StackedParentsProvider(_LazyListJoin(
1081
[self._make_parents_provider_unstacked()],
1082
self._fallback_repositories))
1084
def _make_parents_provider_unstacked(self):
1085
return graph.CallableToParentsProviderAdapter(
1086
self._get_parent_map_no_fallbacks)
1089
def get_known_graph_ancestry(self, revision_ids):
1090
"""Return the known graph for a set of revision ids and their ancestors.
1092
raise NotImplementedError(self.get_known_graph_ancestry)
1094
def get_file_graph(self):
1095
"""Return the graph walker for files."""
1096
raise NotImplementedError(self.get_file_graph)
1098
1773
def get_graph(self, other_repository=None):
1099
1774
"""Return the graph walker for this repository format"""
1100
1775
parents_provider = self._make_parents_provider()
1101
1776
if (other_repository is not None and
1102
1777
not self.has_same_location(other_repository)):
1103
parents_provider = graph.StackedParentsProvider(
1778
parents_provider = graph._StackedParentsProvider(
1104
1779
[parents_provider, other_repository._make_parents_provider()])
1105
1780
return graph.Graph(parents_provider)
1782
def _get_versioned_file_checker(self):
1783
"""Return an object suitable for checking versioned files."""
1784
return _VersionedFileChecker(self)
1786
def revision_ids_to_search_result(self, result_set):
1787
"""Convert a set of revision ids to a graph SearchResult."""
1788
result_parents = set()
1789
for parents in self.get_graph().get_parent_map(
1790
result_set).itervalues():
1791
result_parents.update(parents)
1792
included_keys = result_set.intersection(result_parents)
1793
start_keys = result_set.difference(included_keys)
1794
exclude_keys = result_parents.difference(result_set)
1795
result = graph.SearchResult(start_keys, exclude_keys,
1796
len(result_set), result_set)
1107
1799
@needs_write_lock
1108
1800
def set_make_working_trees(self, new_value):
1109
1801
"""Set the policy flag for making working trees when creating branches.
1220
1871
revision_id.decode('ascii')
1221
1872
except UnicodeDecodeError:
1222
1873
raise errors.NonAsciiRevisionId(method, self)
1875
def revision_graph_can_have_wrong_parents(self):
1876
"""Is it possible for this repository to have a revision graph with
1879
If True, then this repository must also implement
1880
_find_inconsistent_revision_parents so that check and reconcile can
1881
check for inconsistencies before proceeding with other checks that may
1882
depend on the revision index being consistent.
1884
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1887
# remove these delegates a while after bzr 0.15
1888
def __make_delegated(name, from_module):
1889
def _deprecated_repository_forwarder():
1890
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1891
% (name, from_module),
1894
m = __import__(from_module, globals(), locals(), [name])
1896
return getattr(m, name)
1897
except AttributeError:
1898
raise AttributeError('module %s has no name %s'
1900
globals()[name] = _deprecated_repository_forwarder
1903
'AllInOneRepository',
1904
'WeaveMetaDirRepository',
1905
'PreSplitOutRepositoryFormat',
1906
'RepositoryFormat4',
1907
'RepositoryFormat5',
1908
'RepositoryFormat6',
1909
'RepositoryFormat7',
1911
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1915
'RepositoryFormatKnit',
1916
'RepositoryFormatKnit1',
1918
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1921
def install_revision(repository, rev, revision_tree):
1922
"""Install all revision data into a repository."""
1923
install_revisions(repository, [(rev, revision_tree, None)])
1926
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1927
"""Install all revision data into a repository.
1929
Accepts an iterable of revision, tree, signature tuples. The signature
1932
repository.start_write_group()
1934
for n, (revision, revision_tree, signature) in enumerate(iterable):
1935
_install_revision(repository, revision, revision_tree, signature)
1937
pb.update('Transferring revisions', n + 1, num_revisions)
1939
repository.abort_write_group()
1942
repository.commit_write_group()
1945
def _install_revision(repository, rev, revision_tree, signature):
1946
"""Install all revision data into a repository."""
1947
present_parents = []
1949
for p_id in rev.parent_ids:
1950
if repository.has_revision(p_id):
1951
present_parents.append(p_id)
1952
parent_trees[p_id] = repository.revision_tree(p_id)
1954
parent_trees[p_id] = repository.revision_tree(None)
1956
inv = revision_tree.inventory
1957
entries = inv.iter_entries()
1958
# backwards compatibility hack: skip the root id.
1959
if not repository.supports_rich_root():
1960
path, root = entries.next()
1961
if root.revision != rev.revision_id:
1962
raise errors.IncompatibleRevision(repr(repository))
1963
# Add the texts that are not already present
1964
for path, ie in entries:
1965
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1966
repository.get_transaction())
1967
if ie.revision not in w:
1969
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1970
# with InventoryEntry.find_previous_heads(). if it is, then there
1971
# is a latent bug here where the parents may have ancestors of each
1973
for revision, tree in parent_trees.iteritems():
1974
if ie.file_id not in tree:
1976
parent_id = tree.inventory[ie.file_id].revision
1977
if parent_id in text_parents:
1979
text_parents.append(parent_id)
1981
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1982
repository.get_transaction())
1983
lines = revision_tree.get_file(ie.file_id).readlines()
1984
vfile.add_lines(rev.revision_id, text_parents, lines)
1986
# install the inventory
1987
repository.add_inventory(rev.revision_id, inv, present_parents)
1988
except errors.RevisionAlreadyPresent:
1990
if signature is not None:
1991
repository.add_signature_text(rev.revision_id, signature)
1992
repository.add_revision(rev.revision_id, rev, inv)
1225
1995
class MetaDirRepository(Repository):
1226
"""Repositories in the new meta-dir layout.
1228
:ivar _transport: Transport for access to repository control files,
1229
typically pointing to .bzr/repository.
1232
def __init__(self, _format, a_bzrdir, control_files):
1233
super(MetaDirRepository, self).__init__(_format, a_bzrdir, control_files)
1234
self._transport = control_files._transport
1996
"""Repositories in the new meta-dir layout."""
1998
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1999
super(MetaDirRepository, self).__init__(_format,
2005
dir_mode = self.control_files._dir_mode
2006
file_mode = self.control_files._file_mode
1236
2009
def is_shared(self):
1237
2010
"""Return True if this repository is flagged as a shared repository."""
1238
return self._transport.has('shared-storage')
2011
return self.control_files._transport.has('shared-storage')
1240
2013
@needs_write_lock
1241
2014
def set_make_working_trees(self, new_value):
1632
2358
_optimisers = []
1633
2359
"""The available optimised InterRepository types."""
1636
2361
def copy_content(self, revision_id=None):
1637
"""Make a complete copy of the content in self into destination.
1639
This is a destructive operation! Do not use it on existing
1642
:param revision_id: Only copy the content needed to construct
1643
revision_id and its parents.
1646
self.target.set_make_working_trees(self.source.make_working_trees())
1647
except NotImplementedError:
1649
self.target.fetch(self.source, revision_id=revision_id)
1652
def fetch(self, revision_id=None, find_ghosts=False):
2362
raise NotImplementedError(self.copy_content)
2364
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
1653
2365
"""Fetch the content required to construct revision_id.
1655
2367
The content is copied from self.source to self.target.
1657
2369
:param revision_id: if None all content is copied, if NULL_REVISION no
1658
2370
content is copied.
2371
:param pb: optional progress bar to use for progress reports. If not
2372
provided a default one will be created.
2374
Returns the copied revision count and the failed revisions in a tuple:
1661
2377
raise NotImplementedError(self.fetch)
1664
def search_missing_revision_ids(self,
1665
revision_id=symbol_versioning.DEPRECATED_PARAMETER,
1666
find_ghosts=True, revision_ids=None, if_present_ids=None,
1668
"""Return the revision ids that source has that target does not.
1670
:param revision_id: only return revision ids included by this
1672
:param revision_ids: return revision ids included by these
1673
revision_ids. NoSuchRevision will be raised if any of these
1674
revisions are not present.
1675
:param if_present_ids: like revision_ids, but will not cause
1676
NoSuchRevision if any of these are absent, instead they will simply
1677
not be in the result. This is useful for e.g. finding revisions
1678
to fetch for tags, which may reference absent revisions.
1679
:param find_ghosts: If True find missing revisions in deep history
1680
rather than just finding the surface difference.
1681
:param limit: Maximum number of revisions to return, topologically
2379
def _walk_to_common_revisions(self, revision_ids):
2380
"""Walk out from revision_ids in source to revisions target has.
2382
:param revision_ids: The start point for the search.
2383
:return: A set of revision ids.
2385
graph = self.source.get_graph()
2386
missing_revs = set()
2387
# ensure we don't pay silly lookup costs.
2388
revision_ids = frozenset(revision_ids)
2389
searcher = graph._make_breadth_first_searcher(revision_ids)
2390
null_set = frozenset([_mod_revision.NULL_REVISION])
2393
next_revs, ghosts = searcher.next_with_ghosts()
2394
except StopIteration:
2396
if revision_ids.intersection(ghosts):
2397
absent_ids = set(revision_ids.intersection(ghosts))
2398
# If all absent_ids are present in target, no error is needed.
2399
absent_ids.difference_update(
2400
self.target.has_revisions(absent_ids))
2402
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2403
# we don't care about other ghosts as we can't fetch them and
2404
# haven't been asked to.
2405
next_revs = set(next_revs)
2406
# we always have NULL_REVISION present.
2407
have_revs = self.target.has_revisions(next_revs).union(null_set)
2408
missing_revs.update(next_revs - have_revs)
2409
searcher.stop_searching_any(have_revs)
2410
return searcher.get_result()
2412
@deprecated_method(symbol_versioning.one_two)
2414
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2415
"""Return the revision ids that source has that target does not.
2417
These are returned in topological order.
2419
:param revision_id: only return revision ids included by this
2421
:param find_ghosts: If True find missing revisions in deep history
2422
rather than just finding the surface difference.
2424
return list(self.search_missing_revision_ids(
2425
revision_id, find_ghosts).get_keys())
2428
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2429
"""Return the revision ids that source has that target does not.
2431
:param revision_id: only return revision ids included by this
2433
:param find_ghosts: If True find missing revisions in deep history
2434
rather than just finding the surface difference.
1683
2435
:return: A bzrlib.graph.SearchResult.
1685
raise NotImplementedError(self.search_missing_revision_ids)
2437
# stop searching at found target revisions.
2438
if not find_ghosts and revision_id is not None:
2439
return self._walk_to_common_revisions([revision_id])
2440
# generic, possibly worst case, slow code path.
2441
target_ids = set(self.target.all_revision_ids())
2442
if revision_id is not None:
2443
source_ids = self.source.get_ancestry(revision_id)
2444
assert source_ids[0] is None
2447
source_ids = self.source.all_revision_ids()
2448
result_set = set(source_ids).difference(target_ids)
2449
return self.source.revision_ids_to_search_result(result_set)
1688
2452
def _same_model(source, target):
1689
"""True if source and target have the same data representation.
1691
Note: this is always called on the base class; overriding it in a
1692
subclass will have no effect.
1695
InterRepository._assert_same_model(source, target)
1697
except errors.IncompatibleRepositories, e:
2453
"""True if source and target have the same data representation."""
2454
if source.supports_rich_root() != target.supports_rich_root():
1701
def _assert_same_model(source, target):
1702
"""Raise an exception if two repositories do not use the same model.
1704
if source.supports_rich_root() != target.supports_rich_root():
1705
raise errors.IncompatibleRepositories(source, target,
1706
"different rich-root support")
1707
2456
if source._serializer != target._serializer:
1708
raise errors.IncompatibleRepositories(source, target,
1709
"different serializers")
2461
class InterSameDataRepository(InterRepository):
2462
"""Code for converting between repositories that represent the same data.
2464
Data format and model must match for this to work.
2468
def _get_repo_format_to_test(self):
2469
"""Repository format for testing with.
2471
InterSameData can pull from subtree to subtree and from non-subtree to
2472
non-subtree, so we test this with the richest repository format.
2474
from bzrlib.repofmt import knitrepo
2475
return knitrepo.RepositoryFormatKnit3()
2478
def is_compatible(source, target):
2479
return InterRepository._same_model(source, target)
2482
def copy_content(self, revision_id=None):
2483
"""Make a complete copy of the content in self into destination.
2485
This copies both the repository's revision data, and configuration information
2486
such as the make_working_trees setting.
2488
This is a destructive operation! Do not use it on existing
2491
:param revision_id: Only copy the content needed to construct
2492
revision_id and its parents.
2495
self.target.set_make_working_trees(self.source.make_working_trees())
2496
except NotImplementedError:
2498
# but don't bother fetching if we have the needed data now.
2499
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2500
self.target.has_revision(revision_id)):
2502
self.target.fetch(self.source, revision_id=revision_id)
2505
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2506
"""See InterRepository.fetch()."""
2507
from bzrlib.fetch import GenericRepoFetcher
2508
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2509
self.source, self.source._format, self.target,
2510
self.target._format)
2511
f = GenericRepoFetcher(to_repository=self.target,
2512
from_repository=self.source,
2513
last_revision=revision_id,
2514
pb=pb, find_ghosts=find_ghosts)
2515
return f.count_copied, f.failed_revisions
2518
class InterWeaveRepo(InterSameDataRepository):
2519
"""Optimised code paths between Weave based repositories.
2521
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2522
implemented lazy inter-object optimisation.
2526
def _get_repo_format_to_test(self):
2527
from bzrlib.repofmt import weaverepo
2528
return weaverepo.RepositoryFormat7()
2531
def is_compatible(source, target):
2532
"""Be compatible with known Weave formats.
2534
We don't test for the stores being of specific types because that
2535
could lead to confusing results, and there is no need to be
2538
from bzrlib.repofmt.weaverepo import (
2544
return (isinstance(source._format, (RepositoryFormat5,
2546
RepositoryFormat7)) and
2547
isinstance(target._format, (RepositoryFormat5,
2549
RepositoryFormat7)))
2550
except AttributeError:
2554
def copy_content(self, revision_id=None):
2555
"""See InterRepository.copy_content()."""
2556
# weave specific optimised path:
2558
self.target.set_make_working_trees(self.source.make_working_trees())
2559
except (errors.RepositoryUpgradeRequired, NotImplemented):
2561
# FIXME do not peek!
2562
if self.source.control_files._transport.listable():
2563
pb = ui.ui_factory.nested_progress_bar()
2565
self.target.weave_store.copy_all_ids(
2566
self.source.weave_store,
2568
from_transaction=self.source.get_transaction(),
2569
to_transaction=self.target.get_transaction())
2570
pb.update('copying inventory', 0, 1)
2571
self.target.control_weaves.copy_multi(
2572
self.source.control_weaves, ['inventory'],
2573
from_transaction=self.source.get_transaction(),
2574
to_transaction=self.target.get_transaction())
2575
self.target._revision_store.text_store.copy_all_ids(
2576
self.source._revision_store.text_store,
2581
self.target.fetch(self.source, revision_id=revision_id)
2584
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2585
"""See InterRepository.fetch()."""
2586
from bzrlib.fetch import GenericRepoFetcher
2587
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2588
self.source, self.source._format, self.target, self.target._format)
2589
f = GenericRepoFetcher(to_repository=self.target,
2590
from_repository=self.source,
2591
last_revision=revision_id,
2592
pb=pb, find_ghosts=find_ghosts)
2593
return f.count_copied, f.failed_revisions
2596
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2597
"""See InterRepository.missing_revision_ids()."""
2598
# we want all revisions to satisfy revision_id in source.
2599
# but we don't want to stat every file here and there.
2600
# we want then, all revisions other needs to satisfy revision_id
2601
# checked, but not those that we have locally.
2602
# so the first thing is to get a subset of the revisions to
2603
# satisfy revision_id in source, and then eliminate those that
2604
# we do already have.
2605
# this is slow on high latency connection to self, but as as this
2606
# disk format scales terribly for push anyway due to rewriting
2607
# inventory.weave, this is considered acceptable.
2609
if revision_id is not None:
2610
source_ids = self.source.get_ancestry(revision_id)
2611
assert source_ids[0] is None
2614
source_ids = self.source._all_possible_ids()
2615
source_ids_set = set(source_ids)
2616
# source_ids is the worst possible case we may need to pull.
2617
# now we want to filter source_ids against what we actually
2618
# have in target, but don't try to check for existence where we know
2619
# we do not have a revision as that would be pointless.
2620
target_ids = set(self.target._all_possible_ids())
2621
possibly_present_revisions = target_ids.intersection(source_ids_set)
2622
actually_present_revisions = set(
2623
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2624
required_revisions = source_ids_set.difference(actually_present_revisions)
2625
if revision_id is not None:
2626
# we used get_ancestry to determine source_ids then we are assured all
2627
# revisions referenced are present as they are installed in topological order.
2628
# and the tip revision was validated by get_ancestry.
2629
result_set = required_revisions
2631
# if we just grabbed the possibly available ids, then
2632
# we only have an estimate of whats available and need to validate
2633
# that against the revision records.
2635
self.source._eliminate_revisions_not_present(required_revisions))
2636
return self.source.revision_ids_to_search_result(result_set)
2639
class InterKnitRepo(InterSameDataRepository):
2640
"""Optimised code paths between Knit based repositories."""
2643
def _get_repo_format_to_test(self):
2644
from bzrlib.repofmt import knitrepo
2645
return knitrepo.RepositoryFormatKnit1()
2648
def is_compatible(source, target):
2649
"""Be compatible with known Knit formats.
2651
We don't test for the stores being of specific types because that
2652
could lead to confusing results, and there is no need to be
2655
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2657
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2658
isinstance(target._format, RepositoryFormatKnit))
2659
except AttributeError:
2661
return are_knits and InterRepository._same_model(source, target)
2664
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2665
"""See InterRepository.fetch()."""
2666
from bzrlib.fetch import KnitRepoFetcher
2667
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2668
self.source, self.source._format, self.target, self.target._format)
2669
f = KnitRepoFetcher(to_repository=self.target,
2670
from_repository=self.source,
2671
last_revision=revision_id,
2672
pb=pb, find_ghosts=find_ghosts)
2673
return f.count_copied, f.failed_revisions
2676
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2677
"""See InterRepository.missing_revision_ids()."""
2678
if revision_id is not None:
2679
source_ids = self.source.get_ancestry(revision_id)
2680
assert source_ids[0] is None
2683
source_ids = self.source.all_revision_ids()
2684
source_ids_set = set(source_ids)
2685
# source_ids is the worst possible case we may need to pull.
2686
# now we want to filter source_ids against what we actually
2687
# have in target, but don't try to check for existence where we know
2688
# we do not have a revision as that would be pointless.
2689
target_ids = set(self.target.all_revision_ids())
2690
possibly_present_revisions = target_ids.intersection(source_ids_set)
2691
actually_present_revisions = set(
2692
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2693
required_revisions = source_ids_set.difference(actually_present_revisions)
2694
if revision_id is not None:
2695
# we used get_ancestry to determine source_ids then we are assured all
2696
# revisions referenced are present as they are installed in topological order.
2697
# and the tip revision was validated by get_ancestry.
2698
result_set = required_revisions
2700
# if we just grabbed the possibly available ids, then
2701
# we only have an estimate of whats available and need to validate
2702
# that against the revision records.
2704
self.source._eliminate_revisions_not_present(required_revisions))
2705
return self.source.revision_ids_to_search_result(result_set)
2708
class InterPackRepo(InterSameDataRepository):
2709
"""Optimised code paths between Pack based repositories."""
2712
def _get_repo_format_to_test(self):
2713
from bzrlib.repofmt import pack_repo
2714
return pack_repo.RepositoryFormatKnitPack1()
2717
def is_compatible(source, target):
2718
"""Be compatible with known Pack formats.
2720
We don't test for the stores being of specific types because that
2721
could lead to confusing results, and there is no need to be
2724
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2726
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2727
isinstance(target._format, RepositoryFormatPack))
2728
except AttributeError:
2730
return are_packs and InterRepository._same_model(source, target)
2733
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2734
"""See InterRepository.fetch()."""
2735
from bzrlib.repofmt.pack_repo import Packer
2736
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2737
self.source, self.source._format, self.target, self.target._format)
2738
self.count_copied = 0
2739
if revision_id is None:
2741
# everything to do - use pack logic
2742
# to fetch from all packs to one without
2743
# inventory parsing etc, IFF nothing to be copied is in the target.
2745
revision_ids = self.source.all_revision_ids()
2746
# implementing the TODO will involve:
2747
# - detecting when all of a pack is selected
2748
# - avoiding as much as possible pre-selection, so the
2749
# more-core routines such as create_pack_from_packs can filter in
2750
# a just-in-time fashion. (though having a HEADS list on a
2751
# repository might make this a lot easier, because we could
2752
# sensibly detect 'new revisions' without doing a full index scan.
2753
elif _mod_revision.is_null(revision_id):
2758
revision_ids = self.search_missing_revision_ids(revision_id,
2759
find_ghosts=find_ghosts).get_keys()
2760
except errors.NoSuchRevision:
2761
raise errors.InstallFailed([revision_id])
2762
packs = self.source._pack_collection.all_packs()
2763
pack = Packer(self.target._pack_collection, packs, '.fetch',
2764
revision_ids).pack()
2765
if pack is not None:
2766
self.target._pack_collection._save_pack_names()
2767
# Trigger an autopack. This may duplicate effort as we've just done
2768
# a pack creation, but for now it is simpler to think about as
2769
# 'upload data, then repack if needed'.
2770
self.target._pack_collection.autopack()
2771
return (pack.get_revision_count(), [])
2776
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2777
"""See InterRepository.missing_revision_ids().
2779
:param find_ghosts: Find ghosts throughout the ancestry of
2782
if not find_ghosts and revision_id is not None:
2783
return self._walk_to_common_revisions([revision_id])
2784
elif revision_id is not None:
2785
source_ids = self.source.get_ancestry(revision_id)
2786
assert source_ids[0] is None
2789
source_ids = self.source.all_revision_ids()
2790
# source_ids is the worst possible case we may need to pull.
2791
# now we want to filter source_ids against what we actually
2792
# have in target, but don't try to check for existence where we know
2793
# we do not have a revision as that would be pointless.
2794
target_ids = set(self.target.all_revision_ids())
2795
result_set = set(source_ids).difference(target_ids)
2796
return self.source.revision_ids_to_search_result(result_set)
2799
class InterModel1and2(InterRepository):
2802
def _get_repo_format_to_test(self):
2806
def is_compatible(source, target):
2807
if not source.supports_rich_root() and target.supports_rich_root():
2813
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2814
"""See InterRepository.fetch()."""
2815
from bzrlib.fetch import Model1toKnit2Fetcher
2816
f = Model1toKnit2Fetcher(to_repository=self.target,
2817
from_repository=self.source,
2818
last_revision=revision_id,
2819
pb=pb, find_ghosts=find_ghosts)
2820
return f.count_copied, f.failed_revisions
2823
def copy_content(self, revision_id=None):
2824
"""Make a complete copy of the content in self into destination.
2826
This is a destructive operation! Do not use it on existing
2829
:param revision_id: Only copy the content needed to construct
2830
revision_id and its parents.
2833
self.target.set_make_working_trees(self.source.make_working_trees())
2834
except NotImplementedError:
2836
# but don't bother fetching if we have the needed data now.
2837
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2838
self.target.has_revision(revision_id)):
2840
self.target.fetch(self.source, revision_id=revision_id)
2843
class InterKnit1and2(InterKnitRepo):
2846
def _get_repo_format_to_test(self):
2850
def is_compatible(source, target):
2851
"""Be compatible with Knit1 source and Knit3 target"""
2852
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2854
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2855
RepositoryFormatKnit3)
2856
from bzrlib.repofmt.pack_repo import (
2857
RepositoryFormatKnitPack1,
2858
RepositoryFormatKnitPack3,
2859
RepositoryFormatPackDevelopment0,
2860
RepositoryFormatPackDevelopment0Subtree,
2863
RepositoryFormatKnit1,
2864
RepositoryFormatKnitPack1,
2865
RepositoryFormatPackDevelopment0,
2868
RepositoryFormatKnit3,
2869
RepositoryFormatKnitPack3,
2870
RepositoryFormatPackDevelopment0Subtree,
2872
return (isinstance(source._format, nosubtrees) and
2873
isinstance(target._format, subtrees))
2874
except AttributeError:
2878
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2879
"""See InterRepository.fetch()."""
2880
from bzrlib.fetch import Knit1to2Fetcher
2881
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2882
self.source, self.source._format, self.target,
2883
self.target._format)
2884
f = Knit1to2Fetcher(to_repository=self.target,
2885
from_repository=self.source,
2886
last_revision=revision_id,
2887
pb=pb, find_ghosts=find_ghosts)
2888
return f.count_copied, f.failed_revisions
2891
class InterDifferingSerializer(InterKnitRepo):
2894
def _get_repo_format_to_test(self):
2898
def is_compatible(source, target):
2899
"""Be compatible with Knit2 source and Knit3 target"""
2900
if source.supports_rich_root() != target.supports_rich_root():
2902
# Ideally, we'd support fetching if the source had no tree references
2903
# even if it supported them...
2904
if (getattr(source, '_format.supports_tree_reference', False) and
2905
not getattr(target, '_format.supports_tree_reference', False)):
2910
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2911
"""See InterRepository.fetch()."""
2912
revision_ids = self.target.search_missing_revision_ids(self.source,
2913
revision_id, find_ghosts=find_ghosts).get_keys()
2914
revision_ids = tsort.topo_sort(
2915
self.source.get_graph().get_parent_map(revision_ids))
2916
def revisions_iterator():
2917
for current_revision_id in revision_ids:
2918
revision = self.source.get_revision(current_revision_id)
2919
tree = self.source.revision_tree(current_revision_id)
2921
signature = self.source.get_signature_text(
2922
current_revision_id)
2923
except errors.NoSuchRevision:
2925
yield revision, tree, signature
2927
my_pb = ui.ui_factory.nested_progress_bar()
2932
install_revisions(self.target, revisions_iterator(),
2933
len(revision_ids), pb)
2935
if my_pb is not None:
2937
return len(revision_ids), 0
2940
class InterRemoteToOther(InterRepository):
2942
def __init__(self, source, target):
2943
InterRepository.__init__(self, source, target)
2944
self._real_inter = None
2947
def is_compatible(source, target):
2948
if not isinstance(source, remote.RemoteRepository):
2950
# Is source's model compatible with target's model?
2951
source._ensure_real()
2952
real_source = source._real_repository
2953
assert not isinstance(real_source, remote.RemoteRepository), (
2954
"We don't support remote repos backed by remote repos yet.")
2955
return InterRepository._same_model(real_source, target)
2958
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2959
"""See InterRepository.fetch()."""
2960
from bzrlib.fetch import RemoteToOtherFetcher
2961
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2962
self.source, self.target, self.target._format)
2963
# TODO: jam 20070210 This should be an assert, not a translate
2964
revision_id = osutils.safe_revision_id(revision_id)
2965
f = RemoteToOtherFetcher(to_repository=self.target,
2966
from_repository=self.source,
2967
last_revision=revision_id,
2968
pb=pb, find_ghosts=find_ghosts)
2969
return f.count_copied, f.failed_revisions
2972
def _get_repo_format_to_test(self):
2976
class InterOtherToRemote(InterRepository):
2978
def __init__(self, source, target):
2979
InterRepository.__init__(self, source, target)
2980
self._real_inter = None
2983
def is_compatible(source, target):
2984
if isinstance(target, remote.RemoteRepository):
2988
def _ensure_real_inter(self):
2989
if self._real_inter is None:
2990
self.target._ensure_real()
2991
real_target = self.target._real_repository
2992
self._real_inter = InterRepository.get(self.source, real_target)
2994
def copy_content(self, revision_id=None):
2995
self._ensure_real_inter()
2996
self._real_inter.copy_content(revision_id=revision_id)
2998
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2999
self._ensure_real_inter()
3000
self._real_inter.fetch(revision_id=revision_id, pb=pb,
3001
find_ghosts=find_ghosts)
3004
def _get_repo_format_to_test(self):
3008
InterRepository.register_optimiser(InterDifferingSerializer)
3009
InterRepository.register_optimiser(InterSameDataRepository)
3010
InterRepository.register_optimiser(InterWeaveRepo)
3011
InterRepository.register_optimiser(InterKnitRepo)
3012
InterRepository.register_optimiser(InterModel1and2)
3013
InterRepository.register_optimiser(InterKnit1and2)
3014
InterRepository.register_optimiser(InterPackRepo)
3015
InterRepository.register_optimiser(InterRemoteToOther)
3016
InterRepository.register_optimiser(InterOtherToRemote)
1712
3019
class CopyConverter(object):
1713
3020
"""A repository conversion tool which just performs a copy of the content.
1715
3022
This is slow but quite reliable.