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
424
return versionedfile.add_lines_with_ghosts(
425
self._new_revision_id, parents, new_lines,
426
nostore_sha=nostore_sha, random_id=self.random_revid,
427
check_content=False)[0:2]
429
versionedfile.clear_cache()
432
class RootCommitBuilder(CommitBuilder):
433
"""This commitbuilder actually records the root id"""
435
# the root entry gets versioned properly by this builder.
436
_versioned_root = True
438
def _check_root(self, ie, parent_invs, tree):
439
"""Helper for record_entry_contents.
441
:param ie: An entry being added.
442
:param parent_invs: The inventories of the parent revisions of the
444
:param tree: The tree that is being committed.
247
448
######################################################################
251
class Repository(_RelockDebugMixin, controldir.ControlComponent):
451
class Repository(object):
252
452
"""Repository holding history for one or more branches.
254
454
The repository holds and retrieves historical information including
255
455
revisions and file history. It's normally accessed only by the Branch,
256
456
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.
458
The Repository builds on top of Stores and a Transport, which respectively
459
describe the disk data format and the way of accessing the (possibly
262
def abort_write_group(self, suppress_errors=False):
463
# What class to use for a CommitBuilder. Often its simpler to change this
464
# in a Repository class subclass rather than to override
465
# get_commit_builder.
466
_commit_builder_class = CommitBuilder
467
# The search regex used by xml based repositories to determine what things
468
# where changed in a single commit.
469
_file_ids_altered_regex = lazy_regex.lazy_compile(
470
r'file_id="(?P<file_id>[^"]+)"'
471
r'.* revision="(?P<revision_id>[^"]+)"'
474
def abort_write_group(self):
263
475
"""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
477
:seealso: start_write_group.
271
479
if self._write_group is not self.get_transaction():
272
480
# 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)
481
raise errors.BzrError('mismatched lock context and write group.')
482
self._abort_write_group()
290
483
self._write_group = None
292
485
def _abort_write_group(self):
293
486
"""Template method for per-repository write group cleanup.
295
This is called during abort before the write group is considered to be
488
This is called during abort before the write group is considered to be
296
489
finished and should cleanup any internal state accrued during the write
297
490
group. There is no requirement that data handed to the repository be
298
491
*not* made available - this is not a rollback - but neither should any
546
773
:param using: If True, list only branches using this repository.
548
775
if using and not self.is_shared():
549
return self.bzrdir.list_branches()
777
return [self.bzrdir.open_branch()]
778
except errors.NotBranchError:
550
780
class Evaluator(object):
552
782
def __init__(self):
553
783
self.first_call = True
555
def __call__(self, controldir):
556
# On the first call, the parameter is always the controldir
785
def __call__(self, bzrdir):
786
# On the first call, the parameter is always the bzrdir
557
787
# containing the current repo.
558
788
if not self.first_call:
560
repository = controldir.open_repository()
790
repository = bzrdir.open_repository()
561
791
except errors.NoRepositoryPresent:
564
return False, ([], repository)
794
return False, (None, repository)
565
795
self.first_call = False
566
value = (controldir.list_branches(), None)
797
value = (bzrdir.open_branch(), None)
798
except errors.NotBranchError:
567
800
return True, value
570
for branches, repository in controldir.ControlDir.find_bzrdirs(
571
self.user_transport, evaluate=Evaluator()):
572
if branches is not None:
803
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
804
self.bzrdir.root_transport, evaluate=Evaluator()):
805
if branch is not None:
806
branches.append(branch)
574
807
if not using and repository is not None:
575
ret.extend(repository.find_branches())
808
branches.extend(repository.find_branches())
811
def get_data_stream(self, revision_ids):
812
raise NotImplementedError(self.get_data_stream)
814
def get_data_stream_for_search(self, search_result):
815
"""Get a data stream that can be inserted to a repository.
817
:param search_result: A bzrlib.graph.SearchResult selecting the
819
:return: A data stream that can be inserted into a repository using
822
raise NotImplementedError(self.get_data_stream_for_search)
824
def insert_data_stream(self, stream):
825
"""XXX What does this really do?
827
Is it a substitute for fetch?
828
Should it manage its own write group ?
830
for item_key, bytes in stream:
831
if item_key[0] == 'file':
832
(file_id,) = item_key[1:]
833
knit = self.weave_store.get_weave_or_empty(
834
file_id, self.get_transaction())
835
elif item_key == ('inventory',):
836
knit = self.get_inventory_weave()
837
elif item_key == ('revisions',):
838
knit = self._revision_store.get_revision_file(
839
self.get_transaction())
840
elif item_key == ('signatures',):
841
knit = self._revision_store.get_signature_file(
842
self.get_transaction())
844
raise errors.RepositoryDataStreamError(
845
"Unrecognised data stream key '%s'" % (item_key,))
846
decoded_list = bencode.bdecode(bytes)
847
format = decoded_list.pop(0)
850
for version, options, parents, some_bytes in decoded_list:
851
data_list.append((version, options, len(some_bytes), parents))
852
knit_bytes += some_bytes
853
buffer = StringIO(knit_bytes)
854
def reader_func(count):
858
return buffer.read(count)
859
knit.insert_data_stream(
860
(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,
863
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
583
864
"""Return the revision ids that other has that this does not.
585
866
These are returned in topological order.
587
868
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
870
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)
871
revision_id, find_ghosts)
873
@deprecated_method(symbol_versioning.one_two)
875
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
876
"""Return the revision ids that other has that this does not.
878
These are returned in topological order.
880
revision_id: only return revision ids included by revision_id.
882
keys = self.search_missing_revision_ids(
883
other, revision_id, find_ghosts).get_keys()
886
parents = other.get_graph().get_parent_map(keys)
889
return tsort.topo_sort(parents)
828
1063
raise NotImplementedError(self.has_revisions)
1065
return self._revision_store.has_revision_id(revision_id,
1066
self.get_transaction())
830
1068
@needs_read_lock
831
1069
def get_revision(self, revision_id):
832
1070
"""Return the Revision object for a named revision."""
833
1071
return self.get_revisions([revision_id])[0]
835
1074
def get_revision_reconcile(self, revision_id):
836
1075
"""'reconcile' helper routine that allows access to a revision always.
838
1077
This variant of get_revision does not cross check the weave graph
839
1078
against the revision one as get_revision does: but it should only
840
1079
be used by reconcile, or reconcile-alike commands that are correcting
841
1080
or testing the revision graph.
843
raise NotImplementedError(self.get_revision_reconcile)
1082
return self._get_revisions([revision_id])[0]
845
1085
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):
1086
"""Get many revisions at once."""
1087
return self._get_revisions(revision_ids)
1090
def _get_revisions(self, revision_ids):
1091
"""Core work logic to get many revisions without sanity checks."""
1092
for rev_id in revision_ids:
1093
if not rev_id or not isinstance(rev_id, basestring):
1094
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1095
revs = self._revision_store.get_revisions(revision_ids,
1096
self.get_transaction())
1098
assert not isinstance(rev.revision_id, unicode)
1099
for parent_id in rev.parent_ids:
1100
assert not isinstance(parent_id, unicode)
1104
def get_revision_xml(self, revision_id):
1105
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1106
# would have already do it.
1107
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1108
rev = self.get_revision(revision_id)
1109
rev_tmp = StringIO()
1110
# the current serializer..
1111
self._revision_store._serializer.write_revision(rev, rev_tmp)
1113
return rev_tmp.getvalue()
1116
def get_deltas_for_revisions(self, revisions):
854
1117
"""Produce a generator of revision deltas.
856
1119
Note that the input is a sequence of REVISIONS, not revision_ids.
857
1120
Trees will be held in memory until the generator exits.
858
1121
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
1123
required_trees = set()
866
1124
for revision in revisions:
867
1125
required_trees.add(revision.revision_id)
868
1126
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
1127
trees = dict((t.get_revision_id(), t) for
1128
t in self.revision_trees(required_trees))
885
1129
for revision in revisions:
886
1130
if not revision.parent_ids:
887
old_tree = self.revision_tree(_mod_revision.NULL_REVISION)
1131
old_tree = self.revision_tree(None)
889
1133
old_tree = trees[revision.parent_ids[0]]
890
1134
yield trees[revision.revision_id].changes_from(old_tree)
892
1136
@needs_read_lock
893
def get_revision_delta(self, revision_id, specific_fileids=None):
1137
def get_revision_delta(self, revision_id):
894
1138
"""Return the delta for one revision.
896
1140
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
1143
r = self.get_revision(revision_id)
904
return list(self.get_deltas_for_revisions([r],
905
specific_fileids=specific_fileids))[0]
1144
return list(self.get_deltas_for_revisions([r]))[0]
907
1146
@needs_write_lock
908
1147
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
909
1148
signature = gpg_strategy.sign(plaintext)
910
1149
self.add_signature_text(revision_id, signature)
912
1152
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)
1153
self._revision_store.add_revision_signature_text(revision_id,
1155
self.get_transaction())
1157
def find_text_key_references(self):
1158
"""Find the text key references within the repository.
1160
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1161
revision_ids. Each altered file-ids has the exact revision_ids that
1162
altered it listed explicitly.
1163
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1164
to whether they were referred to by the inventory of the
1165
revision_id that they contain. The inventory texts from all present
1166
revision ids are assessed to generate this report.
1168
revision_ids = self.all_revision_ids()
1169
w = self.get_inventory_weave()
1170
pb = ui.ui_factory.nested_progress_bar()
1172
return self._find_text_key_references_from_xml_inventory_lines(
1173
w.iter_lines_added_or_present_in_versions(revision_ids, pb=pb))
1177
def _find_text_key_references_from_xml_inventory_lines(self,
1179
"""Core routine for extracting references to texts from inventories.
1181
This performs the translation of xml lines to revision ids.
1183
:param line_iterator: An iterator of lines, origin_version_id
1184
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1185
to whether they were referred to by the inventory of the
1186
revision_id that they contain. Note that if that revision_id was
1187
not part of the line_iterator's output then False will be given -
1188
even though it may actually refer to that key.
1190
if not self._serializer.support_altered_by_hack:
1191
raise AssertionError(
1192
"_find_text_key_references_from_xml_inventory_lines only "
1193
"supported for branches which store inventory as unnested xml"
1194
", not on %r" % self)
1197
# this code needs to read every new line in every inventory for the
1198
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1199
# not present in one of those inventories is unnecessary but not
1200
# harmful because we are filtering by the revision id marker in the
1201
# inventory lines : we only select file ids altered in one of those
1202
# revisions. We don't need to see all lines in the inventory because
1203
# only those added in an inventory in rev X can contain a revision=X
1205
unescape_revid_cache = {}
1206
unescape_fileid_cache = {}
1208
# jam 20061218 In a big fetch, this handles hundreds of thousands
1209
# of lines, so it has had a lot of inlining and optimizing done.
1210
# Sorry that it is a little bit messy.
1211
# Move several functions to be local variables, since this is a long
1213
search = self._file_ids_altered_regex.search
1214
unescape = _unescape_xml
1215
setdefault = result.setdefault
1216
for line, version_id in line_iterator:
1217
match = search(line)
1220
# One call to match.group() returning multiple items is quite a
1221
# bit faster than 2 calls to match.group() each returning 1
1222
file_id, revision_id = match.group('file_id', 'revision_id')
1224
# Inlining the cache lookups helps a lot when you make 170,000
1225
# lines and 350k ids, versus 8.4 unique ids.
1226
# Using a cache helps in 2 ways:
1227
# 1) Avoids unnecessary decoding calls
1228
# 2) Re-uses cached strings, which helps in future set and
1230
# (2) is enough that removing encoding entirely along with
1231
# the cache (so we are using plain strings) results in no
1232
# performance improvement.
1234
revision_id = unescape_revid_cache[revision_id]
1236
unescaped = unescape(revision_id)
1237
unescape_revid_cache[revision_id] = unescaped
1238
revision_id = unescaped
1240
# Note that unconditionally unescaping means that we deserialise
1241
# every fileid, which for general 'pull' is not great, but we don't
1242
# really want to have some many fulltexts that this matters anyway.
1245
file_id = unescape_fileid_cache[file_id]
1247
unescaped = unescape(file_id)
1248
unescape_fileid_cache[file_id] = unescaped
1251
key = (file_id, revision_id)
1252
setdefault(key, False)
1253
if revision_id == version_id:
1257
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1259
"""Helper routine for fileids_altered_by_revision_ids.
1261
This performs the translation of xml lines to revision ids.
1263
:param line_iterator: An iterator of lines, origin_version_id
1264
:param revision_ids: The revision ids to filter for. This should be a
1265
set or other type which supports efficient __contains__ lookups, as
1266
the revision id from each parsed line will be looked up in the
1267
revision_ids filter.
1268
:return: a dictionary mapping altered file-ids to an iterable of
1269
revision_ids. Each altered file-ids has the exact revision_ids that
1270
altered it listed explicitly.
1273
setdefault = result.setdefault
1274
for file_id, revision_id in \
1275
self._find_text_key_references_from_xml_inventory_lines(
1276
line_iterator).iterkeys():
1277
# once data is all ensured-consistent; then this is
1278
# if revision_id == version_id
1279
if revision_id in revision_ids:
1280
setdefault(file_id, set()).add(revision_id)
1283
def fileids_altered_by_revision_ids(self, revision_ids):
1284
"""Find the file ids and versions affected by revisions.
1286
:param revisions: an iterable containing revision ids.
1287
:return: a dictionary mapping altered file-ids to an iterable of
1288
revision_ids. Each altered file-ids has the exact revision_ids that
1289
altered it listed explicitly.
1291
selected_revision_ids = set(revision_ids)
1292
w = self.get_inventory_weave()
1293
pb = ui.ui_factory.nested_progress_bar()
1295
return self._find_file_ids_from_xml_inventory_lines(
1296
w.iter_lines_added_or_present_in_versions(
1297
selected_revision_ids, pb=pb),
1298
selected_revision_ids)
933
1302
def iter_files_bytes(self, desired_files):
934
1303
"""Iterate through file versions.
941
1310
uniquely identify the file version in the caller's context. (Examples:
942
1311
an index number or a TreeTransform trans_id.)
1313
bytes_iterator is an iterable of bytestrings for the file. The
1314
kind of iterable and length of the bytestrings are unspecified, but for
1315
this implementation, it is a list of lines produced by
1316
VersionedFile.get_lines().
944
1318
: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])
1321
transaction = self.get_transaction()
1322
for file_id, revision_id, callable_data in desired_files:
1324
weave = self.weave_store.get_weave(file_id, transaction)
1325
except errors.NoSuchFile:
1326
raise errors.NoSuchIdInRepository(self, file_id)
1327
yield callable_data, weave.get_lines(revision_id)
1329
def _generate_text_key_index(self, text_key_references=None,
1331
"""Generate a new text key index for the repository.
1333
This is an expensive function that will take considerable time to run.
1335
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1336
list of parents, also text keys. When a given key has no parents,
1337
the parents list will be [NULL_REVISION].
1339
# All revisions, to find inventory parents.
1340
if ancestors is None:
1341
revision_graph = self.get_revision_graph_with_ghosts()
1342
ancestors = revision_graph.get_ancestors()
1343
if text_key_references is None:
1344
text_key_references = self.find_text_key_references()
1345
pb = ui.ui_factory.nested_progress_bar()
1347
return self._do_generate_text_key_index(ancestors,
1348
text_key_references, pb)
1352
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1353
"""Helper for _generate_text_key_index to avoid deep nesting."""
1354
revision_order = tsort.topo_sort(ancestors)
1355
invalid_keys = set()
1357
for revision_id in revision_order:
1358
revision_keys[revision_id] = set()
1359
text_count = len(text_key_references)
1360
# a cache of the text keys to allow reuse; costs a dict of all the
1361
# keys, but saves a 2-tuple for every child of a given key.
1363
for text_key, valid in text_key_references.iteritems():
1365
invalid_keys.add(text_key)
1367
revision_keys[text_key[1]].add(text_key)
1368
text_key_cache[text_key] = text_key
1369
del text_key_references
1371
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1372
NULL_REVISION = _mod_revision.NULL_REVISION
1373
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1374
# too small for large or very branchy trees. However, for 55K path
1375
# trees, it would be easy to use too much memory trivially. Ideally we
1376
# could gauge this by looking at available real memory etc, but this is
1377
# always a tricky proposition.
1378
inventory_cache = lru_cache.LRUCache(10)
1379
batch_size = 10 # should be ~150MB on a 55K path tree
1380
batch_count = len(revision_order) / batch_size + 1
1382
pb.update("Calculating text parents.", processed_texts, text_count)
1383
for offset in xrange(batch_count):
1384
to_query = revision_order[offset * batch_size:(offset + 1) *
1388
for rev_tree in self.revision_trees(to_query):
1389
revision_id = rev_tree.get_revision_id()
1390
parent_ids = ancestors[revision_id]
1391
for text_key in revision_keys[revision_id]:
1392
pb.update("Calculating text parents.", processed_texts)
1393
processed_texts += 1
1394
candidate_parents = []
1395
for parent_id in parent_ids:
1396
parent_text_key = (text_key[0], parent_id)
1398
check_parent = parent_text_key not in \
1399
revision_keys[parent_id]
1401
# the parent parent_id is a ghost:
1402
check_parent = False
1403
# truncate the derived graph against this ghost.
1404
parent_text_key = None
1406
# look at the parent commit details inventories to
1407
# determine possible candidates in the per file graph.
1410
inv = inventory_cache[parent_id]
1412
inv = self.revision_tree(parent_id).inventory
1413
inventory_cache[parent_id] = inv
1414
parent_entry = inv._byid.get(text_key[0], None)
1415
if parent_entry is not None:
1417
text_key[0], parent_entry.revision)
1419
parent_text_key = None
1420
if parent_text_key is not None:
1421
candidate_parents.append(
1422
text_key_cache[parent_text_key])
1423
parent_heads = text_graph.heads(candidate_parents)
1424
new_parents = list(parent_heads)
1425
new_parents.sort(key=lambda x:candidate_parents.index(x))
1426
if new_parents == []:
1427
new_parents = [NULL_REVISION]
1428
text_index[text_key] = new_parents
1430
for text_key in invalid_keys:
1431
text_index[text_key] = [NULL_REVISION]
1434
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1435
"""Get an iterable listing the keys of all the data introduced by a set
1438
The keys will be ordered so that the corresponding items can be safely
1439
fetched and inserted in that order.
1441
:returns: An iterable producing tuples of (knit-kind, file-id,
1442
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1443
'revisions'. file-id is None unless knit-kind is 'file'.
1445
# XXX: it's a bit weird to control the inventory weave caching in this
1446
# generator. Ideally the caching would be done in fetch.py I think. Or
1447
# maybe this generator should explicitly have the contract that it
1448
# should not be iterated until the previously yielded item has been
1451
inv_w = self.get_inventory_weave()
1452
inv_w.enable_cache()
1454
# file ids that changed
1455
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1457
num_file_ids = len(file_ids)
1458
for file_id, altered_versions in file_ids.iteritems():
1459
if _files_pb is not None:
1460
_files_pb.update("fetch texts", count, num_file_ids)
1462
yield ("file", file_id, altered_versions)
1463
# We're done with the files_pb. Note that it finished by the caller,
1464
# just as it was created by the caller.
1468
yield ("inventory", None, revision_ids)
1472
revisions_with_signatures = set()
1473
for rev_id in revision_ids:
1475
self.get_signature_text(rev_id)
1476
except errors.NoSuchRevision:
1480
revisions_with_signatures.add(rev_id)
1482
yield ("signatures", None, revisions_with_signatures)
1485
yield ("revisions", None, revision_ids)
1488
def get_inventory_weave(self):
1489
return self.control_weaves.get_weave('inventory',
1490
self.get_transaction())
1493
def get_inventory(self, revision_id):
1494
"""Get Inventory object by revision id."""
1495
return self.iter_inventories([revision_id]).next()
1497
def iter_inventories(self, revision_ids):
1498
"""Get many inventories by revision_ids.
1500
This will buffer some or all of the texts used in constructing the
1501
inventories in memory, but will only parse a single inventory at a
1504
:return: An iterator of inventories.
1506
assert None not in revision_ids
1507
assert _mod_revision.NULL_REVISION not in revision_ids
1508
return self._iter_inventories(revision_ids)
1510
def _iter_inventories(self, revision_ids):
1511
"""single-document based inventory iteration."""
1512
texts = self.get_inventory_weave().get_texts(revision_ids)
1513
for text, revision_id in zip(texts, revision_ids):
1514
yield self.deserialise_inventory(revision_id, text)
1516
def deserialise_inventory(self, revision_id, xml):
1517
"""Transform the xml into an inventory object.
1519
:param revision_id: The expected revision id of the inventory.
1520
:param xml: A serialised inventory.
1522
result = self._serializer.read_inventory_from_string(xml, revision_id)
1523
if result.revision_id != revision_id:
1524
raise AssertionError('revision id mismatch %s != %s' % (
1525
result.revision_id, revision_id))
1528
def serialise_inventory(self, inv):
1529
return self._serializer.write_inventory_to_string(inv)
1531
def _serialise_inventory_to_lines(self, inv):
1532
return self._serializer.write_inventory_to_lines(inv)
1534
def get_serializer_format(self):
1535
return self._serializer.format_num
1538
def get_inventory_xml(self, revision_id):
1539
"""Get inventory XML as a file object."""
1541
assert isinstance(revision_id, str), type(revision_id)
1542
iw = self.get_inventory_weave()
1543
return iw.get_text(revision_id)
1545
raise errors.HistoryMissing(self, 'inventory', revision_id)
1548
def get_inventory_sha1(self, revision_id):
1549
"""Return the sha1 hash of the inventory entry
1551
return self.get_revision(revision_id).inventory_sha1
1554
def get_revision_graph(self, revision_id=None):
1555
"""Return a dictionary containing the revision graph.
1557
NB: This method should not be used as it accesses the entire graph all
1558
at once, which is much more data than most operations should require.
1560
:param revision_id: The revision_id to get a graph from. If None, then
1561
the entire revision graph is returned. This is a deprecated mode of
1562
operation and will be removed in the future.
1563
:return: a dictionary of revision_id->revision_parents_list.
1565
raise NotImplementedError(self.get_revision_graph)
1568
def get_revision_graph_with_ghosts(self, revision_ids=None):
1569
"""Return a graph of the revisions with ghosts marked as applicable.
1571
:param revision_ids: an iterable of revisions to graph or None for all.
1572
:return: a Graph object with the graph reachable from revision_ids.
1574
if 'evil' in debug.debug_flags:
1576
"get_revision_graph_with_ghosts scales with size of history.")
1577
result = deprecated_graph.Graph()
1578
if not revision_ids:
1579
pending = set(self.all_revision_ids())
1582
pending = set(revision_ids)
1583
# special case NULL_REVISION
1584
if _mod_revision.NULL_REVISION in pending:
1585
pending.remove(_mod_revision.NULL_REVISION)
1586
required = set(pending)
1589
revision_id = pending.pop()
1591
rev = self.get_revision(revision_id)
1592
except errors.NoSuchRevision:
1593
if revision_id in required:
1596
result.add_ghost(revision_id)
1598
for parent_id in rev.parent_ids:
1599
# is this queued or done ?
1600
if (parent_id not in pending and
1601
parent_id not in done):
1603
pending.add(parent_id)
1604
result.add_node(revision_id, rev.parent_ids)
1605
done.add(revision_id)
1608
def _get_history_vf(self):
1609
"""Get a versionedfile whose history graph reflects all revisions.
1611
For weave repositories, this is the inventory weave.
1613
return self.get_inventory_weave()
1615
def iter_reverse_revision_history(self, revision_id):
1616
"""Iterate backwards through revision ids in the lefthand history
1618
:param revision_id: The revision id to start with. All its lefthand
1619
ancestors will be traversed.
1621
if revision_id in (None, _mod_revision.NULL_REVISION):
1623
next_id = revision_id
1624
versionedfile = self._get_history_vf()
1627
parents = versionedfile.get_parents(next_id)
1628
if len(parents) == 0:
1631
next_id = parents[0]
1634
def get_revision_inventory(self, revision_id):
1635
"""Return inventory of a past revision."""
1636
# TODO: Unify this with get_inventory()
1637
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1638
# must be the same as its revision, so this is trivial.
1639
if revision_id is None:
1640
# This does not make sense: if there is no revision,
1641
# then it is the current tree inventory surely ?!
1642
# and thus get_root_id() is something that looks at the last
1643
# commit on the branch, and the get_root_id is an inventory check.
1644
raise NotImplementedError
1645
# return Inventory(self.get_root_id())
1647
return self.get_inventory(revision_id)
984
1650
def is_shared(self):
985
1651
"""Return True if this repository is flagged as a shared repository."""
986
1652
raise NotImplementedError(self.is_shared)
1000
1666
for repositories to maintain loaded indices across multiple locks
1001
1667
by checking inside their implementation of this method to see
1002
1668
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
1669
the disk format being validatable in this manner.
1010
1672
@needs_read_lock
1011
1673
def revision_tree(self, revision_id):
1012
1674
"""Return Tree for a revision on this branch.
1014
`revision_id` may be NULL_REVISION for the empty tree revision.
1676
`revision_id` may be None for the empty tree revision.
1016
raise NotImplementedError(self.revision_tree)
1678
# TODO: refactor this to use an existing revision object
1679
# so we don't need to read it in twice.
1680
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1681
return RevisionTree(self, Inventory(root_id=None),
1682
_mod_revision.NULL_REVISION)
1684
inv = self.get_revision_inventory(revision_id)
1685
return RevisionTree(self, inv, revision_id)
1018
1688
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:'
1689
"""Return Tree for a revision on this branch.
1691
`revision_id` may not be None or 'null:'"""
1692
inventories = self.iter_inventories(revision_ids)
1693
for inv in inventories:
1694
yield RevisionTree(self, inv, inv.revision_id)
1697
def get_ancestry(self, revision_id, topo_sorted=True):
1698
"""Return a list of revision-ids integrated by a revision.
1700
The first element of the list is always None, indicating the origin
1701
revision. This might change when we have history horizons, or
1702
perhaps we should have a new API.
1704
This is topologically sorted.
1024
raise NotImplementedError(self.revision_trees)
1706
if _mod_revision.is_null(revision_id):
1708
if not self.has_revision(revision_id):
1709
raise errors.NoSuchRevision(self, revision_id)
1710
w = self.get_inventory_weave()
1711
candidates = w.get_ancestry(revision_id, topo_sorted)
1712
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1026
def pack(self, hint=None, clean_obsolete_packs=False):
1027
1715
"""Compress the data within the repository.
1029
1717
This operation only makes sense for some repository types. For other
1030
1718
types it should be a no-op that just returns.
1032
1720
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
1721
@needs_write_lock as this is a long running call its reasonable to
1034
1722
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
1726
def print_file(self, file, revision_id):
1727
"""Print `file` to stdout.
1729
FIXME RBC 20060125 as John Meinel points out this is a bad api
1730
- it writes to stdout, it assumes that that is valid etc. Fix
1731
by creating a new more flexible convenience function.
1733
tree = self.revision_tree(revision_id)
1734
# use inventory as it was in that revision
1735
file_id = tree.inventory.path2id(file)
1737
# TODO: jam 20060427 Write a test for this code path
1738
# it had a bug in it, and was raising the wrong
1740
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1741
tree.print_file(file_id)
1047
1743
def get_transaction(self):
1048
1744
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:
1746
def revision_parents(self, revision_id):
1747
return self.get_inventory_weave().parent_names(revision_id)
1749
@deprecated_method(symbol_versioning.one_one)
1750
def get_parents(self, revision_ids):
1751
"""See StackedParentsProvider.get_parents"""
1752
parent_map = self.get_parent_map(revision_ids)
1753
return [parent_map.get(r, None) for r in revision_ids]
1755
def get_parent_map(self, keys):
1756
"""See graph._StackedParentsProvider.get_parent_map"""
1758
for revision_id in keys:
1061
1759
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,)
1760
parent_map[revision_id] = ()
1763
parent_id_list = self.get_revision(revision_id).parent_ids
1764
except errors.NoSuchRevision:
1767
if len(parent_id_list) == 0:
1768
parent_ids = (_mod_revision.NULL_REVISION,)
1770
parent_ids = tuple(parent_id_list)
1771
parent_map[revision_id] = parent_ids
1077
1774
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
1777
def get_graph(self, other_repository=None):
1099
1778
"""Return the graph walker for this repository format"""
1100
1779
parents_provider = self._make_parents_provider()
1101
1780
if (other_repository is not None and
1102
not self.has_same_location(other_repository)):
1103
parents_provider = graph.StackedParentsProvider(
1781
other_repository.bzrdir.transport.base !=
1782
self.bzrdir.transport.base):
1783
parents_provider = graph._StackedParentsProvider(
1104
1784
[parents_provider, other_repository._make_parents_provider()])
1105
1785
return graph.Graph(parents_provider)
1787
def _get_versioned_file_checker(self):
1788
"""Return an object suitable for checking versioned files."""
1789
return _VersionedFileChecker(self)
1791
def revision_ids_to_search_result(self, result_set):
1792
"""Convert a set of revision ids to a graph SearchResult."""
1793
result_parents = set()
1794
for parents in self.get_graph().get_parent_map(
1795
result_set).itervalues():
1796
result_parents.update(parents)
1797
included_keys = result_set.intersection(result_parents)
1798
start_keys = result_set.difference(included_keys)
1799
exclude_keys = result_parents.difference(result_set)
1800
result = graph.SearchResult(start_keys, exclude_keys,
1801
len(result_set), result_set)
1107
1804
@needs_write_lock
1108
1805
def set_make_working_trees(self, new_value):
1109
1806
"""Set the policy flag for making working trees when creating branches.
1220
1876
revision_id.decode('ascii')
1221
1877
except UnicodeDecodeError:
1222
1878
raise errors.NonAsciiRevisionId(method, self)
1880
def revision_graph_can_have_wrong_parents(self):
1881
"""Is it possible for this repository to have a revision graph with
1884
If True, then this repository must also implement
1885
_find_inconsistent_revision_parents so that check and reconcile can
1886
check for inconsistencies before proceeding with other checks that may
1887
depend on the revision index being consistent.
1889
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1892
# remove these delegates a while after bzr 0.15
1893
def __make_delegated(name, from_module):
1894
def _deprecated_repository_forwarder():
1895
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1896
% (name, from_module),
1899
m = __import__(from_module, globals(), locals(), [name])
1901
return getattr(m, name)
1902
except AttributeError:
1903
raise AttributeError('module %s has no name %s'
1905
globals()[name] = _deprecated_repository_forwarder
1908
'AllInOneRepository',
1909
'WeaveMetaDirRepository',
1910
'PreSplitOutRepositoryFormat',
1911
'RepositoryFormat4',
1912
'RepositoryFormat5',
1913
'RepositoryFormat6',
1914
'RepositoryFormat7',
1916
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1920
'RepositoryFormatKnit',
1921
'RepositoryFormatKnit1',
1923
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1926
def install_revision(repository, rev, revision_tree):
1927
"""Install all revision data into a repository."""
1928
install_revisions(repository, [(rev, revision_tree, None)])
1931
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1932
"""Install all revision data into a repository.
1934
Accepts an iterable of revision, tree, signature tuples. The signature
1937
repository.start_write_group()
1939
for n, (revision, revision_tree, signature) in enumerate(iterable):
1940
_install_revision(repository, revision, revision_tree, signature)
1942
pb.update('Transferring revisions', n + 1, num_revisions)
1944
repository.abort_write_group()
1947
repository.commit_write_group()
1950
def _install_revision(repository, rev, revision_tree, signature):
1951
"""Install all revision data into a repository."""
1952
present_parents = []
1954
for p_id in rev.parent_ids:
1955
if repository.has_revision(p_id):
1956
present_parents.append(p_id)
1957
parent_trees[p_id] = repository.revision_tree(p_id)
1959
parent_trees[p_id] = repository.revision_tree(None)
1961
inv = revision_tree.inventory
1962
entries = inv.iter_entries()
1963
# backwards compatibility hack: skip the root id.
1964
if not repository.supports_rich_root():
1965
path, root = entries.next()
1966
if root.revision != rev.revision_id:
1967
raise errors.IncompatibleRevision(repr(repository))
1968
# Add the texts that are not already present
1969
for path, ie in entries:
1970
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1971
repository.get_transaction())
1972
if ie.revision not in w:
1974
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1975
# with InventoryEntry.find_previous_heads(). if it is, then there
1976
# is a latent bug here where the parents may have ancestors of each
1978
for revision, tree in parent_trees.iteritems():
1979
if ie.file_id not in tree:
1981
parent_id = tree.inventory[ie.file_id].revision
1982
if parent_id in text_parents:
1984
text_parents.append(parent_id)
1986
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1987
repository.get_transaction())
1988
lines = revision_tree.get_file(ie.file_id).readlines()
1989
vfile.add_lines(rev.revision_id, text_parents, lines)
1991
# install the inventory
1992
repository.add_inventory(rev.revision_id, inv, present_parents)
1993
except errors.RevisionAlreadyPresent:
1995
if signature is not None:
1996
repository.add_signature_text(rev.revision_id, signature)
1997
repository.add_revision(rev.revision_id, rev, inv)
1225
2000
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
2001
"""Repositories in the new meta-dir layout."""
2003
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
2004
super(MetaDirRepository, self).__init__(_format,
2010
dir_mode = self.control_files._dir_mode
2011
file_mode = self.control_files._file_mode
1236
2014
def is_shared(self):
1237
2015
"""Return True if this repository is flagged as a shared repository."""
1238
return self._transport.has('shared-storage')
2016
return self.control_files._transport.has('shared-storage')
1240
2018
@needs_write_lock
1241
2019
def set_make_working_trees(self, new_value):
1632
2332
_optimisers = []
1633
2333
"""The available optimised InterRepository types."""
1636
2335
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):
2336
raise NotImplementedError(self.copy_content)
2338
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
1653
2339
"""Fetch the content required to construct revision_id.
1655
2341
The content is copied from self.source to self.target.
1657
2343
:param revision_id: if None all content is copied, if NULL_REVISION no
1658
2344
content is copied.
2345
:param pb: optional progress bar to use for progress reports. If not
2346
provided a default one will be created.
2348
Returns the copied revision count and the failed revisions in a tuple:
1661
2351
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
2353
def _walk_to_common_revisions(self, revision_ids):
2354
"""Walk out from revision_ids in source to revisions target has.
2356
:param revision_ids: The start point for the search.
2357
:return: A set of revision ids.
2359
graph = self.source.get_graph()
2360
missing_revs = set()
2361
# ensure we don't pay silly lookup costs.
2362
revision_ids = frozenset(revision_ids)
2363
searcher = graph._make_breadth_first_searcher(revision_ids)
2364
null_set = frozenset([_mod_revision.NULL_REVISION])
2367
next_revs, ghosts = searcher.next_with_ghosts()
2368
except StopIteration:
2370
if revision_ids.intersection(ghosts):
2371
absent_ids = set(revision_ids.intersection(ghosts))
2372
# If all absent_ids are present in target, no error is needed.
2373
absent_ids.difference_update(
2374
self.target.has_revisions(absent_ids))
2376
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2377
# we don't care about other ghosts as we can't fetch them and
2378
# haven't been asked to.
2379
next_revs = set(next_revs)
2380
# we always have NULL_REVISION present.
2381
have_revs = self.target.has_revisions(next_revs).union(null_set)
2382
missing_revs.update(next_revs - have_revs)
2383
searcher.stop_searching_any(have_revs)
2384
return searcher.get_result()
2386
@deprecated_method(symbol_versioning.one_two)
2388
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2389
"""Return the revision ids that source has that target does not.
2391
These are returned in topological order.
2393
:param revision_id: only return revision ids included by this
2395
:param find_ghosts: If True find missing revisions in deep history
2396
rather than just finding the surface difference.
2398
return list(self.search_missing_revision_ids(
2399
revision_id, find_ghosts).get_keys())
2402
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2403
"""Return the revision ids that source has that target does not.
2405
:param revision_id: only return revision ids included by this
2407
:param find_ghosts: If True find missing revisions in deep history
2408
rather than just finding the surface difference.
1683
2409
:return: A bzrlib.graph.SearchResult.
1685
raise NotImplementedError(self.search_missing_revision_ids)
2411
# stop searching at found target revisions.
2412
if not find_ghosts and revision_id is not None:
2413
return self._walk_to_common_revisions([revision_id])
2414
# generic, possibly worst case, slow code path.
2415
target_ids = set(self.target.all_revision_ids())
2416
if revision_id is not None:
2417
source_ids = self.source.get_ancestry(revision_id)
2418
assert source_ids[0] is None
2421
source_ids = self.source.all_revision_ids()
2422
result_set = set(source_ids).difference(target_ids)
2423
return self.source.revision_ids_to_search_result(result_set)
1688
2426
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:
2427
"""True if source and target have the same data representation."""
2428
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
2430
if source._serializer != target._serializer:
1708
raise errors.IncompatibleRepositories(source, target,
1709
"different serializers")
2435
class InterSameDataRepository(InterRepository):
2436
"""Code for converting between repositories that represent the same data.
2438
Data format and model must match for this to work.
2442
def _get_repo_format_to_test(self):
2443
"""Repository format for testing with.
2445
InterSameData can pull from subtree to subtree and from non-subtree to
2446
non-subtree, so we test this with the richest repository format.
2448
from bzrlib.repofmt import knitrepo
2449
return knitrepo.RepositoryFormatKnit3()
2452
def is_compatible(source, target):
2453
return InterRepository._same_model(source, target)
2456
def copy_content(self, revision_id=None):
2457
"""Make a complete copy of the content in self into destination.
2459
This copies both the repository's revision data, and configuration information
2460
such as the make_working_trees setting.
2462
This is a destructive operation! Do not use it on existing
2465
:param revision_id: Only copy the content needed to construct
2466
revision_id and its parents.
2469
self.target.set_make_working_trees(self.source.make_working_trees())
2470
except NotImplementedError:
2472
# but don't bother fetching if we have the needed data now.
2473
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2474
self.target.has_revision(revision_id)):
2476
self.target.fetch(self.source, revision_id=revision_id)
2479
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2480
"""See InterRepository.fetch()."""
2481
from bzrlib.fetch import GenericRepoFetcher
2482
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2483
self.source, self.source._format, self.target,
2484
self.target._format)
2485
f = GenericRepoFetcher(to_repository=self.target,
2486
from_repository=self.source,
2487
last_revision=revision_id,
2488
pb=pb, find_ghosts=find_ghosts)
2489
return f.count_copied, f.failed_revisions
2492
class InterWeaveRepo(InterSameDataRepository):
2493
"""Optimised code paths between Weave based repositories.
2495
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2496
implemented lazy inter-object optimisation.
2500
def _get_repo_format_to_test(self):
2501
from bzrlib.repofmt import weaverepo
2502
return weaverepo.RepositoryFormat7()
2505
def is_compatible(source, target):
2506
"""Be compatible with known Weave formats.
2508
We don't test for the stores being of specific types because that
2509
could lead to confusing results, and there is no need to be
2512
from bzrlib.repofmt.weaverepo import (
2518
return (isinstance(source._format, (RepositoryFormat5,
2520
RepositoryFormat7)) and
2521
isinstance(target._format, (RepositoryFormat5,
2523
RepositoryFormat7)))
2524
except AttributeError:
2528
def copy_content(self, revision_id=None):
2529
"""See InterRepository.copy_content()."""
2530
# weave specific optimised path:
2532
self.target.set_make_working_trees(self.source.make_working_trees())
2533
except NotImplementedError:
2535
# FIXME do not peek!
2536
if self.source.control_files._transport.listable():
2537
pb = ui.ui_factory.nested_progress_bar()
2539
self.target.weave_store.copy_all_ids(
2540
self.source.weave_store,
2542
from_transaction=self.source.get_transaction(),
2543
to_transaction=self.target.get_transaction())
2544
pb.update('copying inventory', 0, 1)
2545
self.target.control_weaves.copy_multi(
2546
self.source.control_weaves, ['inventory'],
2547
from_transaction=self.source.get_transaction(),
2548
to_transaction=self.target.get_transaction())
2549
self.target._revision_store.text_store.copy_all_ids(
2550
self.source._revision_store.text_store,
2555
self.target.fetch(self.source, revision_id=revision_id)
2558
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2559
"""See InterRepository.fetch()."""
2560
from bzrlib.fetch import GenericRepoFetcher
2561
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2562
self.source, self.source._format, self.target, self.target._format)
2563
f = GenericRepoFetcher(to_repository=self.target,
2564
from_repository=self.source,
2565
last_revision=revision_id,
2566
pb=pb, find_ghosts=find_ghosts)
2567
return f.count_copied, f.failed_revisions
2570
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2571
"""See InterRepository.missing_revision_ids()."""
2572
# we want all revisions to satisfy revision_id in source.
2573
# but we don't want to stat every file here and there.
2574
# we want then, all revisions other needs to satisfy revision_id
2575
# checked, but not those that we have locally.
2576
# so the first thing is to get a subset of the revisions to
2577
# satisfy revision_id in source, and then eliminate those that
2578
# we do already have.
2579
# this is slow on high latency connection to self, but as as this
2580
# disk format scales terribly for push anyway due to rewriting
2581
# inventory.weave, this is considered acceptable.
2583
if revision_id is not None:
2584
source_ids = self.source.get_ancestry(revision_id)
2585
assert source_ids[0] is None
2588
source_ids = self.source._all_possible_ids()
2589
source_ids_set = set(source_ids)
2590
# source_ids is the worst possible case we may need to pull.
2591
# now we want to filter source_ids against what we actually
2592
# have in target, but don't try to check for existence where we know
2593
# we do not have a revision as that would be pointless.
2594
target_ids = set(self.target._all_possible_ids())
2595
possibly_present_revisions = target_ids.intersection(source_ids_set)
2596
actually_present_revisions = set(
2597
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2598
required_revisions = source_ids_set.difference(actually_present_revisions)
2599
if revision_id is not None:
2600
# we used get_ancestry to determine source_ids then we are assured all
2601
# revisions referenced are present as they are installed in topological order.
2602
# and the tip revision was validated by get_ancestry.
2603
result_set = required_revisions
2605
# if we just grabbed the possibly available ids, then
2606
# we only have an estimate of whats available and need to validate
2607
# that against the revision records.
2609
self.source._eliminate_revisions_not_present(required_revisions))
2610
return self.source.revision_ids_to_search_result(result_set)
2613
class InterKnitRepo(InterSameDataRepository):
2614
"""Optimised code paths between Knit based repositories."""
2617
def _get_repo_format_to_test(self):
2618
from bzrlib.repofmt import knitrepo
2619
return knitrepo.RepositoryFormatKnit1()
2622
def is_compatible(source, target):
2623
"""Be compatible with known Knit formats.
2625
We don't test for the stores being of specific types because that
2626
could lead to confusing results, and there is no need to be
2629
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2631
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2632
isinstance(target._format, RepositoryFormatKnit))
2633
except AttributeError:
2635
return are_knits and InterRepository._same_model(source, target)
2638
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2639
"""See InterRepository.fetch()."""
2640
from bzrlib.fetch import KnitRepoFetcher
2641
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2642
self.source, self.source._format, self.target, self.target._format)
2643
f = KnitRepoFetcher(to_repository=self.target,
2644
from_repository=self.source,
2645
last_revision=revision_id,
2646
pb=pb, find_ghosts=find_ghosts)
2647
return f.count_copied, f.failed_revisions
2650
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2651
"""See InterRepository.missing_revision_ids()."""
2652
if revision_id is not None:
2653
source_ids = self.source.get_ancestry(revision_id)
2654
assert source_ids[0] is None
2657
source_ids = self.source.all_revision_ids()
2658
source_ids_set = set(source_ids)
2659
# source_ids is the worst possible case we may need to pull.
2660
# now we want to filter source_ids against what we actually
2661
# have in target, but don't try to check for existence where we know
2662
# we do not have a revision as that would be pointless.
2663
target_ids = set(self.target.all_revision_ids())
2664
possibly_present_revisions = target_ids.intersection(source_ids_set)
2665
actually_present_revisions = set(
2666
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2667
required_revisions = source_ids_set.difference(actually_present_revisions)
2668
if revision_id is not None:
2669
# we used get_ancestry to determine source_ids then we are assured all
2670
# revisions referenced are present as they are installed in topological order.
2671
# and the tip revision was validated by get_ancestry.
2672
result_set = required_revisions
2674
# if we just grabbed the possibly available ids, then
2675
# we only have an estimate of whats available and need to validate
2676
# that against the revision records.
2678
self.source._eliminate_revisions_not_present(required_revisions))
2679
return self.source.revision_ids_to_search_result(result_set)
2682
class InterPackRepo(InterSameDataRepository):
2683
"""Optimised code paths between Pack based repositories."""
2686
def _get_repo_format_to_test(self):
2687
from bzrlib.repofmt import pack_repo
2688
return pack_repo.RepositoryFormatKnitPack1()
2691
def is_compatible(source, target):
2692
"""Be compatible with known Pack formats.
2694
We don't test for the stores being of specific types because that
2695
could lead to confusing results, and there is no need to be
2698
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2700
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2701
isinstance(target._format, RepositoryFormatPack))
2702
except AttributeError:
2704
return are_packs and InterRepository._same_model(source, target)
2707
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2708
"""See InterRepository.fetch()."""
2709
from bzrlib.repofmt.pack_repo import Packer
2710
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2711
self.source, self.source._format, self.target, self.target._format)
2712
self.count_copied = 0
2713
if revision_id is None:
2715
# everything to do - use pack logic
2716
# to fetch from all packs to one without
2717
# inventory parsing etc, IFF nothing to be copied is in the target.
2719
revision_ids = self.source.all_revision_ids()
2720
# implementing the TODO will involve:
2721
# - detecting when all of a pack is selected
2722
# - avoiding as much as possible pre-selection, so the
2723
# more-core routines such as create_pack_from_packs can filter in
2724
# a just-in-time fashion. (though having a HEADS list on a
2725
# repository might make this a lot easier, because we could
2726
# sensibly detect 'new revisions' without doing a full index scan.
2727
elif _mod_revision.is_null(revision_id):
2732
revision_ids = self.search_missing_revision_ids(revision_id,
2733
find_ghosts=find_ghosts).get_keys()
2734
except errors.NoSuchRevision:
2735
raise errors.InstallFailed([revision_id])
2736
packs = self.source._pack_collection.all_packs()
2737
pack = Packer(self.target._pack_collection, packs, '.fetch',
2738
revision_ids).pack()
2739
if pack is not None:
2740
self.target._pack_collection._save_pack_names()
2741
# Trigger an autopack. This may duplicate effort as we've just done
2742
# a pack creation, but for now it is simpler to think about as
2743
# 'upload data, then repack if needed'.
2744
self.target._pack_collection.autopack()
2745
return (pack.get_revision_count(), [])
2750
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2751
"""See InterRepository.missing_revision_ids().
2753
:param find_ghosts: Find ghosts throughout the ancestry of
2756
if not find_ghosts and revision_id is not None:
2757
return self._walk_to_common_revisions([revision_id])
2758
elif revision_id is not None:
2759
source_ids = self.source.get_ancestry(revision_id)
2760
assert source_ids[0] is None
2763
source_ids = self.source.all_revision_ids()
2764
# source_ids is the worst possible case we may need to pull.
2765
# now we want to filter source_ids against what we actually
2766
# have in target, but don't try to check for existence where we know
2767
# we do not have a revision as that would be pointless.
2768
target_ids = set(self.target.all_revision_ids())
2769
result_set = set(source_ids).difference(target_ids)
2770
return self.source.revision_ids_to_search_result(result_set)
2773
class InterModel1and2(InterRepository):
2776
def _get_repo_format_to_test(self):
2780
def is_compatible(source, target):
2781
if not source.supports_rich_root() and target.supports_rich_root():
2787
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2788
"""See InterRepository.fetch()."""
2789
from bzrlib.fetch import Model1toKnit2Fetcher
2790
f = Model1toKnit2Fetcher(to_repository=self.target,
2791
from_repository=self.source,
2792
last_revision=revision_id,
2793
pb=pb, find_ghosts=find_ghosts)
2794
return f.count_copied, f.failed_revisions
2797
def copy_content(self, revision_id=None):
2798
"""Make a complete copy of the content in self into destination.
2800
This is a destructive operation! Do not use it on existing
2803
:param revision_id: Only copy the content needed to construct
2804
revision_id and its parents.
2807
self.target.set_make_working_trees(self.source.make_working_trees())
2808
except NotImplementedError:
2810
# but don't bother fetching if we have the needed data now.
2811
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2812
self.target.has_revision(revision_id)):
2814
self.target.fetch(self.source, revision_id=revision_id)
2817
class InterKnit1and2(InterKnitRepo):
2820
def _get_repo_format_to_test(self):
2824
def is_compatible(source, target):
2825
"""Be compatible with Knit1 source and Knit3 target"""
2826
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2828
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2829
RepositoryFormatKnit3)
2830
from bzrlib.repofmt.pack_repo import (RepositoryFormatKnitPack1,
2831
RepositoryFormatKnitPack3)
2832
return (isinstance(source._format,
2833
(RepositoryFormatKnit1, RepositoryFormatKnitPack1)) and
2834
isinstance(target._format,
2835
(RepositoryFormatKnit3, RepositoryFormatKnitPack3))
2837
except AttributeError:
2841
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2842
"""See InterRepository.fetch()."""
2843
from bzrlib.fetch import Knit1to2Fetcher
2844
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2845
self.source, self.source._format, self.target,
2846
self.target._format)
2847
f = Knit1to2Fetcher(to_repository=self.target,
2848
from_repository=self.source,
2849
last_revision=revision_id,
2850
pb=pb, find_ghosts=find_ghosts)
2851
return f.count_copied, f.failed_revisions
2854
class InterDifferingSerializer(InterKnitRepo):
2857
def _get_repo_format_to_test(self):
2861
def is_compatible(source, target):
2862
"""Be compatible with Knit2 source and Knit3 target"""
2863
if source.supports_rich_root() != target.supports_rich_root():
2865
# Ideally, we'd support fetching if the source had no tree references
2866
# even if it supported them...
2867
if (getattr(source, '_format.supports_tree_reference', False) and
2868
not getattr(target, '_format.supports_tree_reference', False)):
2873
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2874
"""See InterRepository.fetch()."""
2875
revision_ids = self.target.search_missing_revision_ids(self.source,
2876
revision_id, find_ghosts=find_ghosts).get_keys()
2877
revision_ids = tsort.topo_sort(
2878
self.source.get_graph().get_parent_map(revision_ids))
2879
def revisions_iterator():
2880
for current_revision_id in revision_ids:
2881
revision = self.source.get_revision(current_revision_id)
2882
tree = self.source.revision_tree(current_revision_id)
2884
signature = self.source.get_signature_text(
2885
current_revision_id)
2886
except errors.NoSuchRevision:
2888
yield revision, tree, signature
2890
my_pb = ui.ui_factory.nested_progress_bar()
2895
install_revisions(self.target, revisions_iterator(),
2896
len(revision_ids), pb)
2898
if my_pb is not None:
2900
return len(revision_ids), 0
2903
class InterRemoteToOther(InterRepository):
2905
def __init__(self, source, target):
2906
InterRepository.__init__(self, source, target)
2907
self._real_inter = None
2910
def is_compatible(source, target):
2911
if not isinstance(source, remote.RemoteRepository):
2913
# Is source's model compatible with target's model?
2914
source._ensure_real()
2915
real_source = source._real_repository
2916
assert not isinstance(real_source, remote.RemoteRepository), (
2917
"We don't support remote repos backed by remote repos yet.")
2918
return InterRepository._same_model(real_source, target)
2921
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2922
"""See InterRepository.fetch()."""
2923
from bzrlib.fetch import RemoteToOtherFetcher
2924
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2925
self.source, self.target, self.target._format)
2926
# TODO: jam 20070210 This should be an assert, not a translate
2927
revision_id = osutils.safe_revision_id(revision_id)
2928
f = RemoteToOtherFetcher(to_repository=self.target,
2929
from_repository=self.source,
2930
last_revision=revision_id,
2931
pb=pb, find_ghosts=find_ghosts)
2932
return f.count_copied, f.failed_revisions
2935
def _get_repo_format_to_test(self):
2939
class InterOtherToRemote(InterRepository):
2941
def __init__(self, source, target):
2942
InterRepository.__init__(self, source, target)
2943
self._real_inter = None
2946
def is_compatible(source, target):
2947
if isinstance(target, remote.RemoteRepository):
2951
def _ensure_real_inter(self):
2952
if self._real_inter is None:
2953
self.target._ensure_real()
2954
real_target = self.target._real_repository
2955
self._real_inter = InterRepository.get(self.source, real_target)
2957
def copy_content(self, revision_id=None):
2958
self._ensure_real_inter()
2959
self._real_inter.copy_content(revision_id=revision_id)
2961
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2962
self._ensure_real_inter()
2963
self._real_inter.fetch(revision_id=revision_id, pb=pb,
2964
find_ghosts=find_ghosts)
2967
def _get_repo_format_to_test(self):
2971
InterRepository.register_optimiser(InterDifferingSerializer)
2972
InterRepository.register_optimiser(InterSameDataRepository)
2973
InterRepository.register_optimiser(InterWeaveRepo)
2974
InterRepository.register_optimiser(InterKnitRepo)
2975
InterRepository.register_optimiser(InterModel1and2)
2976
InterRepository.register_optimiser(InterKnit1and2)
2977
InterRepository.register_optimiser(InterPackRepo)
2978
InterRepository.register_optimiser(InterRemoteToOther)
2979
InterRepository.register_optimiser(InterOtherToRemote)
1712
2982
class CopyConverter(object):
1713
2983
"""A repository conversion tool which just performs a copy of the content.
1715
2985
This is slow but quite reliable.
1721
2991
:param target_format: The format the resulting repository should be.
1723
2993
self.target_format = target_format
1725
2995
def convert(self, repo, pb):
1726
2996
"""Perform the conversion of to_convert, giving feedback via pb.
1728
2998
:param to_convert: The disk object to convert.
1729
2999
:param pb: a progress bar to use for progress information.
1731
pb = ui.ui_factory.nested_progress_bar()
1734
3004
# this is only useful with metadir layouts - separated repo content.
1735
3005
# trigger an assertion if not such
1736
3006
repo._format.get_format_string()
1737
3007
self.repo_dir = repo.bzrdir
1738
pb.update(gettext('Moving repository to repository.backup'))
3008
self.step('Moving repository to repository.backup')
1739
3009
self.repo_dir.transport.move('repository', 'repository.backup')
1740
3010
backup_transport = self.repo_dir.transport.clone('repository.backup')
1741
3011
repo._format.check_conversion_target(self.target_format)
1742
3012
self.source_repo = repo._format.open(self.repo_dir,
1744
3014
_override_transport=backup_transport)
1745
pb.update(gettext('Creating new repository'))
3015
self.step('Creating new repository')
1746
3016
converted = self.target_format.initialize(self.repo_dir,
1747
3017
self.source_repo.is_shared())
1748
3018
converted.lock_write()
1750
pb.update(gettext('Copying content'))
3020
self.step('Copying content into repository.')
1751
3021
self.source_repo.copy_content_into(converted)
1753
3023
converted.unlock()
1754
pb.update(gettext('Deleting old repository content'))
3024
self.step('Deleting old repository content.')
1755
3025
self.repo_dir.transport.delete_tree('repository.backup')
1756
ui.ui_factory.note(gettext('repository converted'))
1760
def _strip_NULL_ghosts(revision_graph):
1761
"""Also don't use this. more compatibility code for unmigrated clients."""
1762
# Filter ghosts, and null:
1763
if _mod_revision.NULL_REVISION in revision_graph:
1764
del revision_graph[_mod_revision.NULL_REVISION]
1765
for key, parents in revision_graph.items():
1766
revision_graph[key] = tuple(parent for parent in parents if parent
1768
return revision_graph
1771
def _iter_for_revno(repo, partial_history_cache, stop_index=None,
1772
stop_revision=None):
1773
"""Extend the partial history to include a given index
1775
If a stop_index is supplied, stop when that index has been reached.
1776
If a stop_revision is supplied, stop when that revision is
1777
encountered. Otherwise, stop when the beginning of history is
1780
:param stop_index: The index which should be present. When it is
1781
present, history extension will stop.
1782
:param stop_revision: The revision id which should be present. When
1783
it is encountered, history extension will stop.
1785
start_revision = partial_history_cache[-1]
1786
graph = repo.get_graph()
1787
iterator = graph.iter_lefthand_ancestry(start_revision,
1788
(_mod_revision.NULL_REVISION,))
3026
self.pb.note('repository converted')
3028
def step(self, message):
3029
"""Update the pb by a step."""
3031
self.pb.update(message, self.count, self.total)
3043
def _unescaper(match, _map=_unescape_map):
3044
code = match.group(1)
1790
# skip the last revision in the list
1793
if (stop_index is not None and
1794
len(partial_history_cache) > stop_index):
1796
if partial_history_cache[-1] == stop_revision:
1798
revision_id = iterator.next()
1799
partial_history_cache.append(revision_id)
1800
except StopIteration:
1805
class _LazyListJoin(object):
1806
"""An iterable yielding the contents of many lists as one list.
1808
Each iterator made from this will reflect the current contents of the lists
1809
at the time the iterator is made.
3048
if not code.startswith('#'):
3050
return unichr(int(code[1:])).encode('utf8')
3056
def _unescape_xml(data):
3057
"""Unescape predefined XML entities in a string of data."""
3059
if _unescape_re is None:
3060
_unescape_re = re.compile('\&([^;]*);')
3061
return _unescape_re.sub(_unescaper, data)
3064
class _VersionedFileChecker(object):
3066
def __init__(self, repository):
3067
self.repository = repository
3068
self.text_index = self.repository._generate_text_key_index()
1811
This is used by Repository's _make_parents_provider implementation so that
1814
pp = repo._make_parents_provider() # uses a list of fallback repos
1815
pp.add_fallback_repository(other_repo) # appends to that list
1816
result = pp.get_parent_map(...)
1817
# The result will include revs from other_repo
1820
def __init__(self, *list_parts):
1821
self.list_parts = list_parts
1825
for list_part in self.list_parts:
1826
full_list.extend(list_part)
1827
return iter(full_list)
1830
return "%s.%s(%s)" % (self.__module__, self.__class__.__name__,
3070
def calculate_file_version_parents(self, revision_id, file_id):
3071
"""Calculate the correct parents for a file version according to
3074
parent_keys = self.text_index[(file_id, revision_id)]
3075
if parent_keys == [_mod_revision.NULL_REVISION]:
3077
# strip the file_id, for the weave api
3078
return tuple([revision_id for file_id, revision_id in parent_keys])
3080
def check_file_version_parents(self, weave, file_id):
3081
"""Check the parents stored in a versioned file are correct.
3083
It also detects file versions that are not referenced by their
3084
corresponding revision's inventory.
3086
:returns: A tuple of (wrong_parents, dangling_file_versions).
3087
wrong_parents is a dict mapping {revision_id: (stored_parents,
3088
correct_parents)} for each revision_id where the stored parents
3089
are not correct. dangling_file_versions is a set of (file_id,
3090
revision_id) tuples for versions that are present in this versioned
3091
file, but not used by the corresponding inventory.
3094
unused_versions = set()
3095
for num, revision_id in enumerate(weave.versions()):
3097
correct_parents = self.calculate_file_version_parents(
3098
revision_id, file_id)
3100
# The version is not part of the used keys.
3101
unused_versions.add(revision_id)
3104
knit_parents = tuple(weave.get_parents(revision_id))
3105
except errors.RevisionNotPresent:
3107
if correct_parents != knit_parents:
3108
wrong_parents[revision_id] = (knit_parents, correct_parents)
3109
return wrong_parents, unused_versions