1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
39
revision as _mod_revision,
44
from bzrlib.bundle import serializer
45
from bzrlib.revisiontree import RevisionTree
46
from bzrlib.store.versioned import VersionedFileStore
47
from bzrlib.store.text import TextStore
48
from bzrlib.testament import Testament
51
from bzrlib.decorators import needs_read_lock, needs_write_lock
52
from bzrlib.inter import InterObject
53
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
54
from bzrlib.symbol_versioning import (
57
from bzrlib.trace import mutter, mutter_callsite, note, warning
60
# Old formats display a warning, but only once
61
_deprecation_warning_done = False
64
class CommitBuilder(object):
65
"""Provides an interface to build up a commit.
67
This allows describing a tree to be committed without needing to
68
know the internals of the format of the repository.
71
# all clients should supply tree roots.
72
record_root_entry = True
73
# the default CommitBuilder does not manage trees whose root is versioned.
74
_versioned_root = False
76
def __init__(self, repository, parents, config, timestamp=None,
77
timezone=None, committer=None, revprops=None,
79
"""Initiate a CommitBuilder.
81
:param repository: Repository to commit to.
82
:param parents: Revision ids of the parents of the new revision.
83
:param config: Configuration to use.
84
:param timestamp: Optional timestamp recorded for commit.
85
:param timezone: Optional timezone for timestamp.
86
:param committer: Optional committer to set for commit.
87
:param revprops: Optional dictionary of revision properties.
88
:param revision_id: Optional revision id.
93
self._committer = self._config.username()
95
assert isinstance(committer, basestring), type(committer)
96
self._committer = committer
98
self.new_inventory = Inventory(None)
99
self._new_revision_id = osutils.safe_revision_id(revision_id)
100
self.parents = parents
101
self.repository = repository
104
if revprops is not None:
105
self._revprops.update(revprops)
107
if timestamp is None:
108
timestamp = time.time()
109
# Restrict resolution to 1ms
110
self._timestamp = round(timestamp, 3)
113
self._timezone = osutils.local_time_offset()
115
self._timezone = int(timezone)
117
self._generate_revision_if_needed()
119
def commit(self, message):
120
"""Make the actual commit.
122
:return: The revision id of the recorded revision.
124
rev = _mod_revision.Revision(
125
timestamp=self._timestamp,
126
timezone=self._timezone,
127
committer=self._committer,
129
inventory_sha1=self.inv_sha1,
130
revision_id=self._new_revision_id,
131
properties=self._revprops)
132
rev.parent_ids = self.parents
133
self.repository.add_revision(self._new_revision_id, rev,
134
self.new_inventory, self._config)
135
self.repository.commit_write_group()
136
return self._new_revision_id
139
"""Abort the commit that is being built.
141
self.repository.abort_write_group()
143
def revision_tree(self):
144
"""Return the tree that was just committed.
146
After calling commit() this can be called to get a RevisionTree
147
representing the newly committed tree. This is preferred to
148
calling Repository.revision_tree() because that may require
149
deserializing the inventory, while we already have a copy in
152
return RevisionTree(self.repository, self.new_inventory,
153
self._new_revision_id)
155
def finish_inventory(self):
156
"""Tell the builder that the inventory is finished."""
157
if self.new_inventory.root is None:
158
symbol_versioning.warn('Root entry should be supplied to'
159
' record_entry_contents, as of bzr 0.10.',
160
DeprecationWarning, stacklevel=2)
161
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
162
self.new_inventory.revision_id = self._new_revision_id
163
self.inv_sha1 = self.repository.add_inventory(
164
self._new_revision_id,
169
def _gen_revision_id(self):
170
"""Return new revision-id."""
171
return generate_ids.gen_revision_id(self._config.username(),
174
def _generate_revision_if_needed(self):
175
"""Create a revision id if None was supplied.
177
If the repository can not support user-specified revision ids
178
they should override this function and raise CannotSetRevisionId
179
if _new_revision_id is not None.
181
:raises: CannotSetRevisionId
183
if self._new_revision_id is None:
184
self._new_revision_id = self._gen_revision_id()
185
self.random_revid = True
187
self.random_revid = False
189
def _check_root(self, ie, parent_invs, tree):
190
"""Helper for record_entry_contents.
192
:param ie: An entry being added.
193
:param parent_invs: The inventories of the parent revisions of the
195
:param tree: The tree that is being committed.
197
if ie.parent_id is not None:
198
# if ie is not root, add a root automatically.
199
symbol_versioning.warn('Root entry should be supplied to'
200
' record_entry_contents, as of bzr 0.10.',
201
DeprecationWarning, stacklevel=2)
202
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
203
'', tree, tree.path_content_summary(''))
205
# In this revision format, root entries have no knit or weave When
206
# serializing out to disk and back in root.revision is always
208
ie.revision = self._new_revision_id
210
def record_entry_contents(self, ie, parent_invs, path, tree,
212
"""Record the content of ie from tree into the commit if needed.
214
Side effect: sets ie.revision when unchanged
216
:param ie: An inventory entry present in the commit.
217
:param parent_invs: The inventories of the parent revisions of the
219
:param path: The path the entry is at in the tree.
220
:param tree: The tree which contains this entry and should be used to
222
:param content_summary: Summary data from the tree about the paths
223
content - stat, length, exec, sha/link target. This is only
224
accessed when the entry has a revision of None - that is when it is
225
a candidate to commit.
226
:return: True if a new version of the entry has been recorded.
227
(Committing a merge where a file was only changed on the other side
228
will not return True.)
230
if self.new_inventory.root is None:
231
self._check_root(ie, parent_invs, tree)
232
if ie.revision is None:
233
kind = content_summary[0]
235
# ie is carried over from a prior commit
237
# XXX: repository specific check for nested tree support goes here - if
238
# the repo doesn't want nested trees we skip it ?
239
if (kind == 'tree-reference' and
240
not self.repository._format.supports_tree_reference):
241
# mismatch between commit builder logic and repository:
242
# this needs the entry creation pushed down into the builder.
243
raise NotImplementedError
244
# transitional assert only, will remove before release.
245
assert ie.kind == kind
246
self.new_inventory.add(ie)
248
# ie.revision is always None if the InventoryEntry is considered
249
# for committing. We may record the previous parents revision if the
250
# content is actually unchanged against a sole head.
251
if ie.revision is not None:
252
return ie.revision == self._new_revision_id and (path != '' or
253
self._versioned_root)
255
# XXX: Friction: parent_candidates should return a list not a dict
256
# so that we don't have to walk the inventories again.
257
parent_candiate_entries = ie.parent_candidates(parent_invs)
258
head_set = self.repository.get_graph().heads(parent_candiate_entries.keys())
260
for inv in parent_invs:
261
if ie.file_id in inv:
262
old_rev = inv[ie.file_id].revision
263
if old_rev in head_set:
264
heads.append(inv[ie.file_id].revision)
265
head_set.remove(inv[ie.file_id].revision)
268
# now we check to see if we need to write a new record to the
270
# We write a new entry unless there is one head to the ancestors, and
271
# the kind-derived content is unchanged.
273
# Cheapest check first: no ancestors, or more the one head in the
274
# ancestors, we write a new node.
278
# There is a single head, look it up for comparison
279
parent_entry = parent_candiate_entries[heads[0]]
280
# if the non-content specific data has changed, we'll be writing a
282
if (parent_entry.parent_id != ie.parent_id or
283
parent_entry.name != ie.name):
285
# now we need to do content specific checks:
287
# if the kind changed the content obviously has
288
if kind != parent_entry.kind:
292
if (# if the file length changed we have to store:
293
parent_entry.text_size != content_summary[1] or
294
# if the exec bit has changed we have to store:
295
parent_entry.executable != content_summary[2]):
297
elif parent_entry.text_sha1 == content_summary[3]:
298
# all meta and content is unchanged (using a hash cache
299
# hit to check the sha)
300
ie.revision = parent_entry.revision
301
ie.text_size = parent_entry.text_size
302
ie.text_sha1 = parent_entry.text_sha1
303
ie.executable = parent_entry.executable
306
# Either there is only a hash change(no hash cache entry,
307
# or same size content change), or there is no change on
309
# There is a race condition when inserting content into the
310
# knit though that can result in different content being
311
# inserted so even though we may have had a hash cache hit
312
# here we still tell the store the hash we would *not*
313
# store a new text on, which means that it can avoid for us
314
# without a race condition and without double-shaing the
316
nostore_sha = parent_entry.text_sha1
320
ie.executable = content_summary[2]
321
lines = tree.get_file(ie.file_id, path).readlines()
322
ie.text_sha1, ie.text_size = self._add_text_to_weave(
323
ie.file_id, lines, heads, nostore_sha)
324
except errors.ExistingContent:
325
# we are not going to store a new file graph node as it turns
326
# out to be unchanged.
327
ie.revision = parent_entry.revision
328
ie.text_size = parent_entry.text_size
329
ie.text_sha1 = parent_entry.text_sha1
330
ie.executable = parent_entry.executable
332
elif kind == 'directory':
334
# all data is meta here, nothing specific to directory, so
336
ie.revision = parent_entry.revision
339
self._add_text_to_weave(ie.file_id, lines, heads, None)
340
elif kind == 'symlink':
341
current_link_target = content_summary[3]
343
# symmlink target is not generic metadata, check if it has
345
if current_link_target != parent_entry.symlink_target:
348
# unchanged, carry over.
349
ie.revision = parent_entry.revision
350
ie.symlink_target = parent_entry.symlink_target
352
ie.symlink_target = current_link_target
354
self._add_text_to_weave(ie.file_id, lines, heads, None)
355
elif kind == 'tree-reference':
357
if content_summary[3] != parent_entry.reference_revision:
360
# unchanged, carry over.
361
ie.reference_revision = parent_entry.reference_revision
362
ie.revision = parent_entry.revision
364
ie.reference_revision = content_summary[3]
366
self._add_text_to_weave(ie.file_id, lines, heads, None)
368
raise NotImplementedError('unknown kind')
369
ie.revision = self._new_revision_id
372
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
373
versionedfile = self.repository.weave_store.get_weave_or_empty(
374
file_id, self.repository.get_transaction())
375
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
376
# than add_lines, and allows committing when a parent is ghosted for
378
# Note: as we read the content directly from the tree, we know its not
379
# been turned into unicode or badly split - but a broken tree
380
# implementation could give us bad output from readlines() so this is
381
# not a guarantee of safety. What would be better is always checking
382
# the content during test suite execution. RBC 20070912
384
return versionedfile.add_lines_with_ghosts(
385
self._new_revision_id, parents, new_lines,
386
nostore_sha=nostore_sha, random_id=self.random_revid,
387
check_content=False)[0:2]
389
versionedfile.clear_cache()
392
class RootCommitBuilder(CommitBuilder):
393
"""This commitbuilder actually records the root id"""
395
# the root entry gets versioned properly by this builder.
396
_versioned_root = True
398
def _check_root(self, ie, parent_invs, tree):
399
"""Helper for record_entry_contents.
401
:param ie: An entry being added.
402
:param parent_invs: The inventories of the parent revisions of the
404
:param tree: The tree that is being committed.
406
# ie must be root for this builder
407
assert ie.parent_id is None
410
######################################################################
413
class Repository(object):
414
"""Repository holding history for one or more branches.
416
The repository holds and retrieves historical information including
417
revisions and file history. It's normally accessed only by the Branch,
418
which views a particular line of development through that history.
420
The Repository builds on top of Stores and a Transport, which respectively
421
describe the disk data format and the way of accessing the (possibly
425
# What class to use for a CommitBuilder. Often its simpler to change this
426
# in a Repository class subclass rather than to override
427
# get_commit_builder.
428
_commit_builder_class = CommitBuilder
429
# The search regex used by xml based repositories to determine what things
430
# where changed in a single commit.
431
_file_ids_altered_regex = lazy_regex.lazy_compile(
432
r'file_id="(?P<file_id>[^"]+)"'
433
r'.* revision="(?P<revision_id>[^"]+)"'
436
def abort_write_group(self):
437
"""Commit the contents accrued within the current write group.
439
:seealso: start_write_group.
441
if self._write_group is not self.get_transaction():
442
# has an unlock or relock occured ?
443
raise errors.BzrError('mismatched lock context and write group.')
444
self._abort_write_group()
445
self._write_group = None
447
def _abort_write_group(self):
448
"""Template method for per-repository write group cleanup.
450
This is called during abort before the write group is considered to be
451
finished and should cleanup any internal state accrued during the write
452
group. There is no requirement that data handed to the repository be
453
*not* made available - this is not a rollback - but neither should any
454
attempt be made to ensure that data added is fully commited. Abort is
455
invoked when an error has occured so futher disk or network operations
456
may not be possible or may error and if possible should not be
461
def add_inventory(self, revision_id, inv, parents):
462
"""Add the inventory inv to the repository as revision_id.
464
:param parents: The revision ids of the parents that revision_id
465
is known to have and are in the repository already.
467
returns the sha1 of the serialized inventory.
469
revision_id = osutils.safe_revision_id(revision_id)
470
_mod_revision.check_not_reserved_id(revision_id)
471
assert inv.revision_id is None or inv.revision_id == revision_id, \
472
"Mismatch between inventory revision" \
473
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
474
assert inv.root is not None
475
inv_lines = self._serialise_inventory_to_lines(inv)
476
inv_vf = self.get_inventory_weave()
477
return self._inventory_add_lines(inv_vf, revision_id, parents,
478
inv_lines, check_content=False)
480
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
482
"""Store lines in inv_vf and return the sha1 of the inventory."""
484
for parent in parents:
486
final_parents.append(parent)
487
return inv_vf.add_lines(revision_id, final_parents, lines,
488
check_content=check_content)[0]
491
def add_revision(self, revision_id, rev, inv=None, config=None):
492
"""Add rev to the revision store as revision_id.
494
:param revision_id: the revision id to use.
495
:param rev: The revision object.
496
:param inv: The inventory for the revision. if None, it will be looked
497
up in the inventory storer
498
:param config: If None no digital signature will be created.
499
If supplied its signature_needed method will be used
500
to determine if a signature should be made.
502
revision_id = osutils.safe_revision_id(revision_id)
503
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
505
_mod_revision.check_not_reserved_id(revision_id)
506
if config is not None and config.signature_needed():
508
inv = self.get_inventory(revision_id)
509
plaintext = Testament(rev, inv).as_short_text()
510
self.store_revision_signature(
511
gpg.GPGStrategy(config), plaintext, revision_id)
512
if not revision_id in self.get_inventory_weave():
514
raise errors.WeaveRevisionNotPresent(revision_id,
515
self.get_inventory_weave())
517
# yes, this is not suitable for adding with ghosts.
518
self.add_inventory(revision_id, inv, rev.parent_ids)
519
self._revision_store.add_revision(rev, self.get_transaction())
521
def _add_revision_text(self, revision_id, text):
522
revision = self._revision_store._serializer.read_revision_from_string(
524
self._revision_store._add_revision(revision, StringIO(text),
525
self.get_transaction())
527
def all_revision_ids(self):
528
"""Returns a list of all the revision ids in the repository.
530
This is deprecated because code should generally work on the graph
531
reachable from a particular revision, and ignore any other revisions
532
that might be present. There is no direct replacement method.
534
if 'evil' in debug.debug_flags:
535
mutter_callsite(2, "all_revision_ids is linear with history.")
536
return self._all_revision_ids()
538
def _all_revision_ids(self):
539
"""Returns a list of all the revision ids in the repository.
541
These are in as much topological order as the underlying store can
544
raise NotImplementedError(self._all_revision_ids)
546
def break_lock(self):
547
"""Break a lock if one is present from another instance.
549
Uses the ui factory to ask for confirmation if the lock may be from
552
self.control_files.break_lock()
555
def _eliminate_revisions_not_present(self, revision_ids):
556
"""Check every revision id in revision_ids to see if we have it.
558
Returns a set of the present revisions.
561
for id in revision_ids:
562
if self.has_revision(id):
567
def create(a_bzrdir):
568
"""Construct the current default format repository in a_bzrdir."""
569
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
571
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
572
"""instantiate a Repository.
574
:param _format: The format of the repository on disk.
575
:param a_bzrdir: The BzrDir of the repository.
577
In the future we will have a single api for all stores for
578
getting file texts, inventories and revisions, then
579
this construct will accept instances of those things.
581
super(Repository, self).__init__()
582
self._format = _format
583
# the following are part of the public API for Repository:
584
self.bzrdir = a_bzrdir
585
self.control_files = control_files
586
self._revision_store = _revision_store
587
# backwards compatibility
588
self.weave_store = text_store
590
self._reconcile_does_inventory_gc = True
591
# not right yet - should be more semantically clear ?
593
self.control_store = control_store
594
self.control_weaves = control_store
595
# TODO: make sure to construct the right store classes, etc, depending
596
# on whether escaping is required.
597
self._warn_if_deprecated()
598
self._write_group = None
601
return '%s(%r)' % (self.__class__.__name__,
602
self.bzrdir.transport.base)
604
def has_same_location(self, other):
605
"""Returns a boolean indicating if this repository is at the same
606
location as another repository.
608
This might return False even when two repository objects are accessing
609
the same physical repository via different URLs.
611
if self.__class__ is not other.__class__:
613
return (self.control_files._transport.base ==
614
other.control_files._transport.base)
616
def is_in_write_group(self):
617
"""Return True if there is an open write group.
619
:seealso: start_write_group.
621
return self._write_group is not None
624
return self.control_files.is_locked()
626
def lock_write(self, token=None):
627
"""Lock this repository for writing.
629
This causes caching within the repository obejct to start accumlating
630
data during reads, and allows a 'write_group' to be obtained. Write
631
groups must be used for actual data insertion.
633
:param token: if this is already locked, then lock_write will fail
634
unless the token matches the existing lock.
635
:returns: a token if this instance supports tokens, otherwise None.
636
:raises TokenLockingNotSupported: when a token is given but this
637
instance doesn't support using token locks.
638
:raises MismatchedToken: if the specified token doesn't match the token
639
of the existing lock.
640
:seealso: start_write_group.
642
A token should be passed in if you know that you have locked the object
643
some other way, and need to synchronise this object's state with that
646
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
648
result = self.control_files.lock_write(token=token)
653
self.control_files.lock_read()
656
def get_physical_lock_status(self):
657
return self.control_files.get_physical_lock_status()
659
def leave_lock_in_place(self):
660
"""Tell this repository not to release the physical lock when this
663
If lock_write doesn't return a token, then this method is not supported.
665
self.control_files.leave_in_place()
667
def dont_leave_lock_in_place(self):
668
"""Tell this repository to release the physical lock when this
669
object is unlocked, even if it didn't originally acquire it.
671
If lock_write doesn't return a token, then this method is not supported.
673
self.control_files.dont_leave_in_place()
676
def gather_stats(self, revid=None, committers=None):
677
"""Gather statistics from a revision id.
679
:param revid: The revision id to gather statistics from, if None, then
680
no revision specific statistics are gathered.
681
:param committers: Optional parameter controlling whether to grab
682
a count of committers from the revision specific statistics.
683
:return: A dictionary of statistics. Currently this contains:
684
committers: The number of committers if requested.
685
firstrev: A tuple with timestamp, timezone for the penultimate left
686
most ancestor of revid, if revid is not the NULL_REVISION.
687
latestrev: A tuple with timestamp, timezone for revid, if revid is
688
not the NULL_REVISION.
689
revisions: The total revision count in the repository.
690
size: An estimate disk size of the repository in bytes.
693
if revid and committers:
694
result['committers'] = 0
695
if revid and revid != _mod_revision.NULL_REVISION:
697
all_committers = set()
698
revisions = self.get_ancestry(revid)
699
# pop the leading None
701
first_revision = None
703
# ignore the revisions in the middle - just grab first and last
704
revisions = revisions[0], revisions[-1]
705
for revision in self.get_revisions(revisions):
706
if not first_revision:
707
first_revision = revision
709
all_committers.add(revision.committer)
710
last_revision = revision
712
result['committers'] = len(all_committers)
713
result['firstrev'] = (first_revision.timestamp,
714
first_revision.timezone)
715
result['latestrev'] = (last_revision.timestamp,
716
last_revision.timezone)
718
# now gather global repository information
719
if self.bzrdir.root_transport.listable():
720
c, t = self._revision_store.total_size(self.get_transaction())
721
result['revisions'] = c
726
def missing_revision_ids(self, other, revision_id=None):
727
"""Return the revision ids that other has that this does not.
729
These are returned in topological order.
731
revision_id: only return revision ids included by revision_id.
733
revision_id = osutils.safe_revision_id(revision_id)
734
return InterRepository.get(other, self).missing_revision_ids(revision_id)
738
"""Open the repository rooted at base.
740
For instance, if the repository is at URL/.bzr/repository,
741
Repository.open(URL) -> a Repository instance.
743
control = bzrdir.BzrDir.open(base)
744
return control.open_repository()
746
def copy_content_into(self, destination, revision_id=None):
747
"""Make a complete copy of the content in self into destination.
749
This is a destructive operation! Do not use it on existing
752
revision_id = osutils.safe_revision_id(revision_id)
753
return InterRepository.get(self, destination).copy_content(revision_id)
755
def commit_write_group(self):
756
"""Commit the contents accrued within the current write group.
758
:seealso: start_write_group.
760
if self._write_group is not self.get_transaction():
761
# has an unlock or relock occured ?
762
raise errors.BzrError('mismatched lock context and write group.')
763
self._commit_write_group()
764
self._write_group = None
766
def _commit_write_group(self):
767
"""Template method for per-repository write group cleanup.
769
This is called before the write group is considered to be
770
finished and should ensure that all data handed to the repository
771
for writing during the write group is safely committed (to the
772
extent possible considering file system caching etc).
775
def fetch(self, source, revision_id=None, pb=None):
776
"""Fetch the content required to construct revision_id from source.
778
If revision_id is None all content is copied.
780
revision_id = osutils.safe_revision_id(revision_id)
781
inter = InterRepository.get(source, self)
783
return inter.fetch(revision_id=revision_id, pb=pb)
784
except NotImplementedError:
785
raise errors.IncompatibleRepositories(source, self)
787
def create_bundle(self, target, base, fileobj, format=None):
788
return serializer.write_bundle(self, target, base, fileobj, format)
790
def get_commit_builder(self, branch, parents, config, timestamp=None,
791
timezone=None, committer=None, revprops=None,
793
"""Obtain a CommitBuilder for this repository.
795
:param branch: Branch to commit to.
796
:param parents: Revision ids of the parents of the new revision.
797
:param config: Configuration to use.
798
:param timestamp: Optional timestamp recorded for commit.
799
:param timezone: Optional timezone for timestamp.
800
:param committer: Optional committer to set for commit.
801
:param revprops: Optional dictionary of revision properties.
802
:param revision_id: Optional revision id.
804
revision_id = osutils.safe_revision_id(revision_id)
805
result = self._commit_builder_class(self, parents, config,
806
timestamp, timezone, committer, revprops, revision_id)
807
self.start_write_group()
811
if (self.control_files._lock_count == 1 and
812
self.control_files._lock_mode == 'w'):
813
if self._write_group is not None:
814
raise errors.BzrError(
815
'Must end write groups before releasing write locks.')
816
self.control_files.unlock()
819
def clone(self, a_bzrdir, revision_id=None):
820
"""Clone this repository into a_bzrdir using the current format.
822
Currently no check is made that the format of this repository and
823
the bzrdir format are compatible. FIXME RBC 20060201.
825
:return: The newly created destination repository.
827
# TODO: deprecate after 0.16; cloning this with all its settings is
828
# probably not very useful -- mbp 20070423
829
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
830
self.copy_content_into(dest_repo, revision_id)
833
def start_write_group(self):
834
"""Start a write group in the repository.
836
Write groups are used by repositories which do not have a 1:1 mapping
837
between file ids and backend store to manage the insertion of data from
838
both fetch and commit operations.
840
A write lock is required around the start_write_group/commit_write_group
841
for the support of lock-requiring repository formats.
843
One can only insert data into a repository inside a write group.
847
if not self.is_locked() or self.control_files._lock_mode != 'w':
848
raise errors.NotWriteLocked(self)
849
if self._write_group:
850
raise errors.BzrError('already in a write group')
851
self._start_write_group()
852
# so we can detect unlock/relock - the write group is now entered.
853
self._write_group = self.get_transaction()
855
def _start_write_group(self):
856
"""Template method for per-repository write group startup.
858
This is called before the write group is considered to be
863
def sprout(self, to_bzrdir, revision_id=None):
864
"""Create a descendent repository for new development.
866
Unlike clone, this does not copy the settings of the repository.
868
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
869
dest_repo.fetch(self, revision_id=revision_id)
872
def _create_sprouting_repo(self, a_bzrdir, shared):
873
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
874
# use target default format.
875
dest_repo = a_bzrdir.create_repository()
877
# Most control formats need the repository to be specifically
878
# created, but on some old all-in-one formats it's not needed
880
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
881
except errors.UninitializableFormat:
882
dest_repo = a_bzrdir.open_repository()
886
def has_revision(self, revision_id):
887
"""True if this repository has a copy of the revision."""
888
if 'evil' in debug.debug_flags:
889
mutter_callsite(2, "has_revision is a LBYL symptom.")
890
revision_id = osutils.safe_revision_id(revision_id)
891
return self._revision_store.has_revision_id(revision_id,
892
self.get_transaction())
895
def get_revision(self, revision_id):
896
"""Return the Revision object for a named revision."""
897
return self.get_revisions([revision_id])[0]
900
def get_revision_reconcile(self, revision_id):
901
"""'reconcile' helper routine that allows access to a revision always.
903
This variant of get_revision does not cross check the weave graph
904
against the revision one as get_revision does: but it should only
905
be used by reconcile, or reconcile-alike commands that are correcting
906
or testing the revision graph.
908
return self._get_revisions([revision_id])[0]
911
def get_revisions(self, revision_ids):
912
"""Get many revisions at once."""
913
return self._get_revisions(revision_ids)
916
def _get_revisions(self, revision_ids):
917
"""Core work logic to get many revisions without sanity checks."""
918
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
919
for rev_id in revision_ids:
920
if not rev_id or not isinstance(rev_id, basestring):
921
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
922
revs = self._revision_store.get_revisions(revision_ids,
923
self.get_transaction())
925
assert not isinstance(rev.revision_id, unicode)
926
for parent_id in rev.parent_ids:
927
assert not isinstance(parent_id, unicode)
931
def get_revision_xml(self, revision_id):
932
# TODO: jam 20070210 This shouldn't be necessary since get_revision
933
# would have already do it.
934
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
935
revision_id = osutils.safe_revision_id(revision_id)
936
rev = self.get_revision(revision_id)
938
# the current serializer..
939
self._revision_store._serializer.write_revision(rev, rev_tmp)
941
return rev_tmp.getvalue()
944
def get_deltas_for_revisions(self, revisions):
945
"""Produce a generator of revision deltas.
947
Note that the input is a sequence of REVISIONS, not revision_ids.
948
Trees will be held in memory until the generator exits.
949
Each delta is relative to the revision's lefthand predecessor.
951
required_trees = set()
952
for revision in revisions:
953
required_trees.add(revision.revision_id)
954
required_trees.update(revision.parent_ids[:1])
955
trees = dict((t.get_revision_id(), t) for
956
t in self.revision_trees(required_trees))
957
for revision in revisions:
958
if not revision.parent_ids:
959
old_tree = self.revision_tree(None)
961
old_tree = trees[revision.parent_ids[0]]
962
yield trees[revision.revision_id].changes_from(old_tree)
965
def get_revision_delta(self, revision_id):
966
"""Return the delta for one revision.
968
The delta is relative to the left-hand predecessor of the
971
r = self.get_revision(revision_id)
972
return list(self.get_deltas_for_revisions([r]))[0]
975
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
976
revision_id = osutils.safe_revision_id(revision_id)
977
signature = gpg_strategy.sign(plaintext)
978
self._revision_store.add_revision_signature_text(revision_id,
980
self.get_transaction())
982
def fileids_altered_by_revision_ids(self, revision_ids):
983
"""Find the file ids and versions affected by revisions.
985
:param revisions: an iterable containing revision ids.
986
:return: a dictionary mapping altered file-ids to an iterable of
987
revision_ids. Each altered file-ids has the exact revision_ids that
988
altered it listed explicitly.
990
assert self._serializer.support_altered_by_hack, \
991
("fileids_altered_by_revision_ids only supported for branches "
992
"which store inventory as unnested xml, not on %r" % self)
993
selected_revision_ids = set(osutils.safe_revision_id(r)
994
for r in revision_ids)
995
w = self.get_inventory_weave()
998
# this code needs to read every new line in every inventory for the
999
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1000
# not present in one of those inventories is unnecessary but not
1001
# harmful because we are filtering by the revision id marker in the
1002
# inventory lines : we only select file ids altered in one of those
1003
# revisions. We don't need to see all lines in the inventory because
1004
# only those added in an inventory in rev X can contain a revision=X
1006
unescape_revid_cache = {}
1007
unescape_fileid_cache = {}
1009
# jam 20061218 In a big fetch, this handles hundreds of thousands
1010
# of lines, so it has had a lot of inlining and optimizing done.
1011
# Sorry that it is a little bit messy.
1012
# Move several functions to be local variables, since this is a long
1014
search = self._file_ids_altered_regex.search
1015
unescape = _unescape_xml
1016
setdefault = result.setdefault
1017
pb = ui.ui_factory.nested_progress_bar()
1019
for line in w.iter_lines_added_or_present_in_versions(
1020
selected_revision_ids, pb=pb):
1021
match = search(line)
1024
# One call to match.group() returning multiple items is quite a
1025
# bit faster than 2 calls to match.group() each returning 1
1026
file_id, revision_id = match.group('file_id', 'revision_id')
1028
# Inlining the cache lookups helps a lot when you make 170,000
1029
# lines and 350k ids, versus 8.4 unique ids.
1030
# Using a cache helps in 2 ways:
1031
# 1) Avoids unnecessary decoding calls
1032
# 2) Re-uses cached strings, which helps in future set and
1034
# (2) is enough that removing encoding entirely along with
1035
# the cache (so we are using plain strings) results in no
1036
# performance improvement.
1038
revision_id = unescape_revid_cache[revision_id]
1040
unescaped = unescape(revision_id)
1041
unescape_revid_cache[revision_id] = unescaped
1042
revision_id = unescaped
1044
if revision_id in selected_revision_ids:
1046
file_id = unescape_fileid_cache[file_id]
1048
unescaped = unescape(file_id)
1049
unescape_fileid_cache[file_id] = unescaped
1051
setdefault(file_id, set()).add(revision_id)
1056
def iter_files_bytes(self, desired_files):
1057
"""Iterate through file versions.
1059
Files will not necessarily be returned in the order they occur in
1060
desired_files. No specific order is guaranteed.
1062
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1063
value supplied by the caller as part of desired_files. It should
1064
uniquely identify the file version in the caller's context. (Examples:
1065
an index number or a TreeTransform trans_id.)
1067
bytes_iterator is an iterable of bytestrings for the file. The
1068
kind of iterable and length of the bytestrings are unspecified, but for
1069
this implementation, it is a list of lines produced by
1070
VersionedFile.get_lines().
1072
:param desired_files: a list of (file_id, revision_id, identifier)
1075
transaction = self.get_transaction()
1076
for file_id, revision_id, callable_data in desired_files:
1078
weave = self.weave_store.get_weave(file_id, transaction)
1079
except errors.NoSuchFile:
1080
raise errors.NoSuchIdInRepository(self, file_id)
1081
yield callable_data, weave.get_lines(revision_id)
1083
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1084
"""Get an iterable listing the keys of all the data introduced by a set
1087
The keys will be ordered so that the corresponding items can be safely
1088
fetched and inserted in that order.
1090
:returns: An iterable producing tuples of (knit-kind, file-id,
1091
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1092
'revisions'. file-id is None unless knit-kind is 'file'.
1094
# XXX: it's a bit weird to control the inventory weave caching in this
1095
# generator. Ideally the caching would be done in fetch.py I think. Or
1096
# maybe this generator should explicitly have the contract that it
1097
# should not be iterated until the previously yielded item has been
1099
inv_w = self.get_inventory_weave()
1100
inv_w.enable_cache()
1102
# file ids that changed
1103
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1105
num_file_ids = len(file_ids)
1106
for file_id, altered_versions in file_ids.iteritems():
1107
if _files_pb is not None:
1108
_files_pb.update("fetch texts", count, num_file_ids)
1110
yield ("file", file_id, altered_versions)
1111
# We're done with the files_pb. Note that it finished by the caller,
1112
# just as it was created by the caller.
1116
yield ("inventory", None, revision_ids)
1120
revisions_with_signatures = set()
1121
for rev_id in revision_ids:
1123
self.get_signature_text(rev_id)
1124
except errors.NoSuchRevision:
1128
revisions_with_signatures.add(rev_id)
1129
yield ("signatures", None, revisions_with_signatures)
1132
yield ("revisions", None, revision_ids)
1135
def get_inventory_weave(self):
1136
return self.control_weaves.get_weave('inventory',
1137
self.get_transaction())
1140
def get_inventory(self, revision_id):
1141
"""Get Inventory object by hash."""
1142
# TODO: jam 20070210 Technically we don't need to sanitize, since all
1143
# called functions must sanitize.
1144
revision_id = osutils.safe_revision_id(revision_id)
1145
return self.deserialise_inventory(
1146
revision_id, self.get_inventory_xml(revision_id))
1148
def deserialise_inventory(self, revision_id, xml):
1149
"""Transform the xml into an inventory object.
1151
:param revision_id: The expected revision id of the inventory.
1152
:param xml: A serialised inventory.
1154
revision_id = osutils.safe_revision_id(revision_id)
1155
result = self._serializer.read_inventory_from_string(xml)
1156
result.root.revision = revision_id
1159
def serialise_inventory(self, inv):
1160
return self._serializer.write_inventory_to_string(inv)
1162
def _serialise_inventory_to_lines(self, inv):
1163
return self._serializer.write_inventory_to_lines(inv)
1165
def get_serializer_format(self):
1166
return self._serializer.format_num
1169
def get_inventory_xml(self, revision_id):
1170
"""Get inventory XML as a file object."""
1171
revision_id = osutils.safe_revision_id(revision_id)
1173
assert isinstance(revision_id, str), type(revision_id)
1174
iw = self.get_inventory_weave()
1175
return iw.get_text(revision_id)
1177
raise errors.HistoryMissing(self, 'inventory', revision_id)
1180
def get_inventory_sha1(self, revision_id):
1181
"""Return the sha1 hash of the inventory entry
1183
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
1184
revision_id = osutils.safe_revision_id(revision_id)
1185
return self.get_revision(revision_id).inventory_sha1
1188
def get_revision_graph(self, revision_id=None):
1189
"""Return a dictionary containing the revision graph.
1191
NB: This method should not be used as it accesses the entire graph all
1192
at once, which is much more data than most operations should require.
1194
:param revision_id: The revision_id to get a graph from. If None, then
1195
the entire revision graph is returned. This is a deprecated mode of
1196
operation and will be removed in the future.
1197
:return: a dictionary of revision_id->revision_parents_list.
1199
raise NotImplementedError(self.get_revision_graph)
1202
def get_revision_graph_with_ghosts(self, revision_ids=None):
1203
"""Return a graph of the revisions with ghosts marked as applicable.
1205
:param revision_ids: an iterable of revisions to graph or None for all.
1206
:return: a Graph object with the graph reachable from revision_ids.
1208
if 'evil' in debug.debug_flags:
1210
"get_revision_graph_with_ghosts scales with size of history.")
1211
result = deprecated_graph.Graph()
1212
if not revision_ids:
1213
pending = set(self.all_revision_ids())
1216
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
1217
# special case NULL_REVISION
1218
if _mod_revision.NULL_REVISION in pending:
1219
pending.remove(_mod_revision.NULL_REVISION)
1220
required = set(pending)
1223
revision_id = pending.pop()
1225
rev = self.get_revision(revision_id)
1226
except errors.NoSuchRevision:
1227
if revision_id in required:
1230
result.add_ghost(revision_id)
1232
for parent_id in rev.parent_ids:
1233
# is this queued or done ?
1234
if (parent_id not in pending and
1235
parent_id not in done):
1237
pending.add(parent_id)
1238
result.add_node(revision_id, rev.parent_ids)
1239
done.add(revision_id)
1242
def _get_history_vf(self):
1243
"""Get a versionedfile whose history graph reflects all revisions.
1245
For weave repositories, this is the inventory weave.
1247
return self.get_inventory_weave()
1249
def iter_reverse_revision_history(self, revision_id):
1250
"""Iterate backwards through revision ids in the lefthand history
1252
:param revision_id: The revision id to start with. All its lefthand
1253
ancestors will be traversed.
1255
revision_id = osutils.safe_revision_id(revision_id)
1256
if revision_id in (None, _mod_revision.NULL_REVISION):
1258
next_id = revision_id
1259
versionedfile = self._get_history_vf()
1262
parents = versionedfile.get_parents(next_id)
1263
if len(parents) == 0:
1266
next_id = parents[0]
1269
def get_revision_inventory(self, revision_id):
1270
"""Return inventory of a past revision."""
1271
# TODO: Unify this with get_inventory()
1272
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1273
# must be the same as its revision, so this is trivial.
1274
if revision_id is None:
1275
# This does not make sense: if there is no revision,
1276
# then it is the current tree inventory surely ?!
1277
# and thus get_root_id() is something that looks at the last
1278
# commit on the branch, and the get_root_id is an inventory check.
1279
raise NotImplementedError
1280
# return Inventory(self.get_root_id())
1282
return self.get_inventory(revision_id)
1285
def is_shared(self):
1286
"""Return True if this repository is flagged as a shared repository."""
1287
raise NotImplementedError(self.is_shared)
1290
def reconcile(self, other=None, thorough=False):
1291
"""Reconcile this repository."""
1292
from bzrlib.reconcile import RepoReconciler
1293
reconciler = RepoReconciler(self, thorough=thorough)
1294
reconciler.reconcile()
1297
def _refresh_data(self):
1298
"""Helper called from lock_* to ensure coherency with disk.
1300
The default implementation does nothing; it is however possible
1301
for repositories to maintain loaded indices across multiple locks
1302
by checking inside their implementation of this method to see
1303
whether their indices are still valid. This depends of course on
1304
the disk format being validatable in this manner.
1308
def revision_tree(self, revision_id):
1309
"""Return Tree for a revision on this branch.
1311
`revision_id` may be None for the empty tree revision.
1313
# TODO: refactor this to use an existing revision object
1314
# so we don't need to read it in twice.
1315
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1316
return RevisionTree(self, Inventory(root_id=None),
1317
_mod_revision.NULL_REVISION)
1319
revision_id = osutils.safe_revision_id(revision_id)
1320
inv = self.get_revision_inventory(revision_id)
1321
return RevisionTree(self, inv, revision_id)
1324
def revision_trees(self, revision_ids):
1325
"""Return Tree for a revision on this branch.
1327
`revision_id` may not be None or 'null:'"""
1328
assert None not in revision_ids
1329
assert _mod_revision.NULL_REVISION not in revision_ids
1330
texts = self.get_inventory_weave().get_texts(revision_ids)
1331
for text, revision_id in zip(texts, revision_ids):
1332
inv = self.deserialise_inventory(revision_id, text)
1333
yield RevisionTree(self, inv, revision_id)
1336
def get_ancestry(self, revision_id, topo_sorted=True):
1337
"""Return a list of revision-ids integrated by a revision.
1339
The first element of the list is always None, indicating the origin
1340
revision. This might change when we have history horizons, or
1341
perhaps we should have a new API.
1343
This is topologically sorted.
1345
if _mod_revision.is_null(revision_id):
1347
revision_id = osutils.safe_revision_id(revision_id)
1348
if not self.has_revision(revision_id):
1349
raise errors.NoSuchRevision(self, revision_id)
1350
w = self.get_inventory_weave()
1351
candidates = w.get_ancestry(revision_id, topo_sorted)
1352
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1355
"""Compress the data within the repository.
1357
This operation only makes sense for some repository types. For other
1358
types it should be a no-op that just returns.
1360
This stub method does not require a lock, but subclasses should use
1361
@needs_write_lock as this is a long running call its reasonable to
1362
implicitly lock for the user.
1366
def print_file(self, file, revision_id):
1367
"""Print `file` to stdout.
1369
FIXME RBC 20060125 as John Meinel points out this is a bad api
1370
- it writes to stdout, it assumes that that is valid etc. Fix
1371
by creating a new more flexible convenience function.
1373
revision_id = osutils.safe_revision_id(revision_id)
1374
tree = self.revision_tree(revision_id)
1375
# use inventory as it was in that revision
1376
file_id = tree.inventory.path2id(file)
1378
# TODO: jam 20060427 Write a test for this code path
1379
# it had a bug in it, and was raising the wrong
1381
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1382
tree.print_file(file_id)
1384
def get_transaction(self):
1385
return self.control_files.get_transaction()
1387
def revision_parents(self, revision_id):
1388
revision_id = osutils.safe_revision_id(revision_id)
1389
return self.get_inventory_weave().parent_names(revision_id)
1391
def get_parents(self, revision_ids):
1392
"""See StackedParentsProvider.get_parents"""
1394
for revision_id in revision_ids:
1395
if revision_id == _mod_revision.NULL_REVISION:
1399
parents = self.get_revision(revision_id).parent_ids
1400
except errors.NoSuchRevision:
1403
if len(parents) == 0:
1404
parents = [_mod_revision.NULL_REVISION]
1405
parents_list.append(parents)
1408
def _make_parents_provider(self):
1411
def get_graph(self, other_repository=None):
1412
"""Return the graph walker for this repository format"""
1413
parents_provider = self._make_parents_provider()
1414
if (other_repository is not None and
1415
other_repository.bzrdir.transport.base !=
1416
self.bzrdir.transport.base):
1417
parents_provider = graph._StackedParentsProvider(
1418
[parents_provider, other_repository._make_parents_provider()])
1419
return graph.Graph(parents_provider)
1422
def set_make_working_trees(self, new_value):
1423
"""Set the policy flag for making working trees when creating branches.
1425
This only applies to branches that use this repository.
1427
The default is 'True'.
1428
:param new_value: True to restore the default, False to disable making
1431
raise NotImplementedError(self.set_make_working_trees)
1433
def make_working_trees(self):
1434
"""Returns the policy for making working trees on new branches."""
1435
raise NotImplementedError(self.make_working_trees)
1438
def sign_revision(self, revision_id, gpg_strategy):
1439
revision_id = osutils.safe_revision_id(revision_id)
1440
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1441
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1444
def has_signature_for_revision_id(self, revision_id):
1445
"""Query for a revision signature for revision_id in the repository."""
1446
revision_id = osutils.safe_revision_id(revision_id)
1447
return self._revision_store.has_signature(revision_id,
1448
self.get_transaction())
1451
def get_signature_text(self, revision_id):
1452
"""Return the text for a signature."""
1453
revision_id = osutils.safe_revision_id(revision_id)
1454
return self._revision_store.get_signature_text(revision_id,
1455
self.get_transaction())
1458
def check(self, revision_ids):
1459
"""Check consistency of all history of given revision_ids.
1461
Different repository implementations should override _check().
1463
:param revision_ids: A non-empty list of revision_ids whose ancestry
1464
will be checked. Typically the last revision_id of a branch.
1466
if not revision_ids:
1467
raise ValueError("revision_ids must be non-empty in %s.check"
1469
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
1470
return self._check(revision_ids)
1472
def _check(self, revision_ids):
1473
result = check.Check(self)
1477
def _warn_if_deprecated(self):
1478
global _deprecation_warning_done
1479
if _deprecation_warning_done:
1481
_deprecation_warning_done = True
1482
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1483
% (self._format, self.bzrdir.transport.base))
1485
def supports_rich_root(self):
1486
return self._format.rich_root_data
1488
def _check_ascii_revisionid(self, revision_id, method):
1489
"""Private helper for ascii-only repositories."""
1490
# weave repositories refuse to store revisionids that are non-ascii.
1491
if revision_id is not None:
1492
# weaves require ascii revision ids.
1493
if isinstance(revision_id, unicode):
1495
revision_id.encode('ascii')
1496
except UnicodeEncodeError:
1497
raise errors.NonAsciiRevisionId(method, self)
1500
revision_id.decode('ascii')
1501
except UnicodeDecodeError:
1502
raise errors.NonAsciiRevisionId(method, self)
1506
# remove these delegates a while after bzr 0.15
1507
def __make_delegated(name, from_module):
1508
def _deprecated_repository_forwarder():
1509
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1510
% (name, from_module),
1513
m = __import__(from_module, globals(), locals(), [name])
1515
return getattr(m, name)
1516
except AttributeError:
1517
raise AttributeError('module %s has no name %s'
1519
globals()[name] = _deprecated_repository_forwarder
1522
'AllInOneRepository',
1523
'WeaveMetaDirRepository',
1524
'PreSplitOutRepositoryFormat',
1525
'RepositoryFormat4',
1526
'RepositoryFormat5',
1527
'RepositoryFormat6',
1528
'RepositoryFormat7',
1530
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1534
'RepositoryFormatKnit',
1535
'RepositoryFormatKnit1',
1537
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1540
def install_revision(repository, rev, revision_tree):
1541
"""Install all revision data into a repository."""
1542
present_parents = []
1544
for p_id in rev.parent_ids:
1545
if repository.has_revision(p_id):
1546
present_parents.append(p_id)
1547
parent_trees[p_id] = repository.revision_tree(p_id)
1549
parent_trees[p_id] = repository.revision_tree(None)
1551
inv = revision_tree.inventory
1552
entries = inv.iter_entries()
1553
# backwards compatibility hack: skip the root id.
1554
if not repository.supports_rich_root():
1555
path, root = entries.next()
1556
if root.revision != rev.revision_id:
1557
raise errors.IncompatibleRevision(repr(repository))
1558
# Add the texts that are not already present
1559
for path, ie in entries:
1560
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1561
repository.get_transaction())
1562
if ie.revision not in w:
1564
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1565
# with InventoryEntry.find_previous_heads(). if it is, then there
1566
# is a latent bug here where the parents may have ancestors of each
1568
for revision, tree in parent_trees.iteritems():
1569
if ie.file_id not in tree:
1571
parent_id = tree.inventory[ie.file_id].revision
1572
if parent_id in text_parents:
1574
text_parents.append(parent_id)
1576
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1577
repository.get_transaction())
1578
lines = revision_tree.get_file(ie.file_id).readlines()
1579
vfile.add_lines(rev.revision_id, text_parents, lines)
1581
# install the inventory
1582
repository.add_inventory(rev.revision_id, inv, present_parents)
1583
except errors.RevisionAlreadyPresent:
1585
repository.add_revision(rev.revision_id, rev, inv)
1588
class MetaDirRepository(Repository):
1589
"""Repositories in the new meta-dir layout."""
1591
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1592
super(MetaDirRepository, self).__init__(_format,
1598
dir_mode = self.control_files._dir_mode
1599
file_mode = self.control_files._file_mode
1602
def is_shared(self):
1603
"""Return True if this repository is flagged as a shared repository."""
1604
return self.control_files._transport.has('shared-storage')
1607
def set_make_working_trees(self, new_value):
1608
"""Set the policy flag for making working trees when creating branches.
1610
This only applies to branches that use this repository.
1612
The default is 'True'.
1613
:param new_value: True to restore the default, False to disable making
1618
self.control_files._transport.delete('no-working-trees')
1619
except errors.NoSuchFile:
1622
self.control_files.put_utf8('no-working-trees', '')
1624
def make_working_trees(self):
1625
"""Returns the policy for making working trees on new branches."""
1626
return not self.control_files._transport.has('no-working-trees')
1629
class RepositoryFormatRegistry(registry.Registry):
1630
"""Registry of RepositoryFormats.
1633
def get(self, format_string):
1634
r = registry.Registry.get(self, format_string)
1640
format_registry = RepositoryFormatRegistry()
1641
"""Registry of formats, indexed by their identifying format string.
1643
This can contain either format instances themselves, or classes/factories that
1644
can be called to obtain one.
1648
#####################################################################
1649
# Repository Formats
1651
class RepositoryFormat(object):
1652
"""A repository format.
1654
Formats provide three things:
1655
* An initialization routine to construct repository data on disk.
1656
* a format string which is used when the BzrDir supports versioned
1658
* an open routine which returns a Repository instance.
1660
Formats are placed in an dict by their format string for reference
1661
during opening. These should be subclasses of RepositoryFormat
1664
Once a format is deprecated, just deprecate the initialize and open
1665
methods on the format class. Do not deprecate the object, as the
1666
object will be created every system load.
1668
Common instance attributes:
1669
_matchingbzrdir - the bzrdir format that the repository format was
1670
originally written to work with. This can be used if manually
1671
constructing a bzrdir and repository, or more commonly for test suite
1676
return "<%s>" % self.__class__.__name__
1678
def __eq__(self, other):
1679
# format objects are generally stateless
1680
return isinstance(other, self.__class__)
1682
def __ne__(self, other):
1683
return not self == other
1686
def find_format(klass, a_bzrdir):
1687
"""Return the format for the repository object in a_bzrdir.
1689
This is used by bzr native formats that have a "format" file in
1690
the repository. Other methods may be used by different types of
1694
transport = a_bzrdir.get_repository_transport(None)
1695
format_string = transport.get("format").read()
1696
return format_registry.get(format_string)
1697
except errors.NoSuchFile:
1698
raise errors.NoRepositoryPresent(a_bzrdir)
1700
raise errors.UnknownFormatError(format=format_string)
1703
def register_format(klass, format):
1704
format_registry.register(format.get_format_string(), format)
1707
def unregister_format(klass, format):
1708
format_registry.remove(format.get_format_string())
1711
def get_default_format(klass):
1712
"""Return the current default format."""
1713
from bzrlib import bzrdir
1714
return bzrdir.format_registry.make_bzrdir('default').repository_format
1716
def _get_control_store(self, repo_transport, control_files):
1717
"""Return the control store for this repository."""
1718
raise NotImplementedError(self._get_control_store)
1720
def get_format_string(self):
1721
"""Return the ASCII format string that identifies this format.
1723
Note that in pre format ?? repositories the format string is
1724
not permitted nor written to disk.
1726
raise NotImplementedError(self.get_format_string)
1728
def get_format_description(self):
1729
"""Return the short description for this format."""
1730
raise NotImplementedError(self.get_format_description)
1732
def _get_revision_store(self, repo_transport, control_files):
1733
"""Return the revision store object for this a_bzrdir."""
1734
raise NotImplementedError(self._get_revision_store)
1736
def _get_text_rev_store(self,
1743
"""Common logic for getting a revision store for a repository.
1745
see self._get_revision_store for the subclass-overridable method to
1746
get the store for a repository.
1748
from bzrlib.store.revision.text import TextRevisionStore
1749
dir_mode = control_files._dir_mode
1750
file_mode = control_files._file_mode
1751
text_store = TextStore(transport.clone(name),
1753
compressed=compressed,
1755
file_mode=file_mode)
1756
_revision_store = TextRevisionStore(text_store, serializer)
1757
return _revision_store
1759
# TODO: this shouldn't be in the base class, it's specific to things that
1760
# use weaves or knits -- mbp 20070207
1761
def _get_versioned_file_store(self,
1766
versionedfile_class=None,
1767
versionedfile_kwargs={},
1769
if versionedfile_class is None:
1770
versionedfile_class = self._versionedfile_class
1771
weave_transport = control_files._transport.clone(name)
1772
dir_mode = control_files._dir_mode
1773
file_mode = control_files._file_mode
1774
return VersionedFileStore(weave_transport, prefixed=prefixed,
1776
file_mode=file_mode,
1777
versionedfile_class=versionedfile_class,
1778
versionedfile_kwargs=versionedfile_kwargs,
1781
def initialize(self, a_bzrdir, shared=False):
1782
"""Initialize a repository of this format in a_bzrdir.
1784
:param a_bzrdir: The bzrdir to put the new repository in it.
1785
:param shared: The repository should be initialized as a sharable one.
1786
:returns: The new repository object.
1788
This may raise UninitializableFormat if shared repository are not
1789
compatible the a_bzrdir.
1791
raise NotImplementedError(self.initialize)
1793
def is_supported(self):
1794
"""Is this format supported?
1796
Supported formats must be initializable and openable.
1797
Unsupported formats may not support initialization or committing or
1798
some other features depending on the reason for not being supported.
1802
def check_conversion_target(self, target_format):
1803
raise NotImplementedError(self.check_conversion_target)
1805
def open(self, a_bzrdir, _found=False):
1806
"""Return an instance of this format for the bzrdir a_bzrdir.
1808
_found is a private parameter, do not use it.
1810
raise NotImplementedError(self.open)
1813
class MetaDirRepositoryFormat(RepositoryFormat):
1814
"""Common base class for the new repositories using the metadir layout."""
1816
rich_root_data = False
1817
supports_tree_reference = False
1818
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1821
super(MetaDirRepositoryFormat, self).__init__()
1823
def _create_control_files(self, a_bzrdir):
1824
"""Create the required files and the initial control_files object."""
1825
# FIXME: RBC 20060125 don't peek under the covers
1826
# NB: no need to escape relative paths that are url safe.
1827
repository_transport = a_bzrdir.get_repository_transport(self)
1828
control_files = lockable_files.LockableFiles(repository_transport,
1829
'lock', lockdir.LockDir)
1830
control_files.create_lock()
1831
return control_files
1833
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1834
"""Upload the initial blank content."""
1835
control_files = self._create_control_files(a_bzrdir)
1836
control_files.lock_write()
1838
control_files._transport.mkdir_multi(dirs,
1839
mode=control_files._dir_mode)
1840
for file, content in files:
1841
control_files.put(file, content)
1842
for file, content in utf8_files:
1843
control_files.put_utf8(file, content)
1845
control_files.put_utf8('shared-storage', '')
1847
control_files.unlock()
1850
# formats which have no format string are not discoverable
1851
# and not independently creatable, so are not registered. They're
1852
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1853
# needed, it's constructed directly by the BzrDir. Non-native formats where
1854
# the repository is not separately opened are similar.
1856
format_registry.register_lazy(
1857
'Bazaar-NG Repository format 7',
1858
'bzrlib.repofmt.weaverepo',
1861
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1862
# default control directory format
1864
format_registry.register_lazy(
1865
'Bazaar-NG Knit Repository Format 1',
1866
'bzrlib.repofmt.knitrepo',
1867
'RepositoryFormatKnit1',
1869
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1871
format_registry.register_lazy(
1872
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1873
'bzrlib.repofmt.knitrepo',
1874
'RepositoryFormatKnit3',
1878
class InterRepository(InterObject):
1879
"""This class represents operations taking place between two repositories.
1881
Its instances have methods like copy_content and fetch, and contain
1882
references to the source and target repositories these operations can be
1885
Often we will provide convenience methods on 'repository' which carry out
1886
operations with another repository - they will always forward to
1887
InterRepository.get(other).method_name(parameters).
1891
"""The available optimised InterRepository types."""
1893
def copy_content(self, revision_id=None):
1894
raise NotImplementedError(self.copy_content)
1896
def fetch(self, revision_id=None, pb=None):
1897
"""Fetch the content required to construct revision_id.
1899
The content is copied from self.source to self.target.
1901
:param revision_id: if None all content is copied, if NULL_REVISION no
1903
:param pb: optional progress bar to use for progress reports. If not
1904
provided a default one will be created.
1906
Returns the copied revision count and the failed revisions in a tuple:
1909
raise NotImplementedError(self.fetch)
1912
def missing_revision_ids(self, revision_id=None):
1913
"""Return the revision ids that source has that target does not.
1915
These are returned in topological order.
1917
:param revision_id: only return revision ids included by this
1920
# generic, possibly worst case, slow code path.
1921
target_ids = set(self.target.all_revision_ids())
1922
if revision_id is not None:
1923
# TODO: jam 20070210 InterRepository is internal enough that it
1924
# should assume revision_ids are already utf-8
1925
revision_id = osutils.safe_revision_id(revision_id)
1926
source_ids = self.source.get_ancestry(revision_id)
1927
assert source_ids[0] is None
1930
source_ids = self.source.all_revision_ids()
1931
result_set = set(source_ids).difference(target_ids)
1932
# this may look like a no-op: its not. It preserves the ordering
1933
# other_ids had while only returning the members from other_ids
1934
# that we've decided we need.
1935
return [rev_id for rev_id in source_ids if rev_id in result_set]
1938
class InterSameDataRepository(InterRepository):
1939
"""Code for converting between repositories that represent the same data.
1941
Data format and model must match for this to work.
1945
def _get_repo_format_to_test(self):
1946
"""Repository format for testing with.
1948
InterSameData can pull from subtree to subtree and from non-subtree to
1949
non-subtree, so we test this with the richest repository format.
1951
from bzrlib.repofmt import knitrepo
1952
return knitrepo.RepositoryFormatKnit3()
1955
def is_compatible(source, target):
1956
if source.supports_rich_root() != target.supports_rich_root():
1958
if source._serializer != target._serializer:
1963
def copy_content(self, revision_id=None):
1964
"""Make a complete copy of the content in self into destination.
1966
This copies both the repository's revision data, and configuration information
1967
such as the make_working_trees setting.
1969
This is a destructive operation! Do not use it on existing
1972
:param revision_id: Only copy the content needed to construct
1973
revision_id and its parents.
1976
self.target.set_make_working_trees(self.source.make_working_trees())
1977
except NotImplementedError:
1979
# TODO: jam 20070210 This is fairly internal, so we should probably
1980
# just assert that revision_id is not unicode.
1981
revision_id = osutils.safe_revision_id(revision_id)
1982
# but don't bother fetching if we have the needed data now.
1983
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1984
self.target.has_revision(revision_id)):
1986
self.target.fetch(self.source, revision_id=revision_id)
1989
def fetch(self, revision_id=None, pb=None):
1990
"""See InterRepository.fetch()."""
1991
from bzrlib.fetch import GenericRepoFetcher
1992
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1993
self.source, self.source._format, self.target,
1994
self.target._format)
1995
# TODO: jam 20070210 This should be an assert, not a translate
1996
revision_id = osutils.safe_revision_id(revision_id)
1997
f = GenericRepoFetcher(to_repository=self.target,
1998
from_repository=self.source,
1999
last_revision=revision_id,
2001
return f.count_copied, f.failed_revisions
2004
class InterWeaveRepo(InterSameDataRepository):
2005
"""Optimised code paths between Weave based repositories.
2007
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2008
implemented lazy inter-object optimisation.
2012
def _get_repo_format_to_test(self):
2013
from bzrlib.repofmt import weaverepo
2014
return weaverepo.RepositoryFormat7()
2017
def is_compatible(source, target):
2018
"""Be compatible with known Weave formats.
2020
We don't test for the stores being of specific types because that
2021
could lead to confusing results, and there is no need to be
2024
from bzrlib.repofmt.weaverepo import (
2030
return (isinstance(source._format, (RepositoryFormat5,
2032
RepositoryFormat7)) and
2033
isinstance(target._format, (RepositoryFormat5,
2035
RepositoryFormat7)))
2036
except AttributeError:
2040
def copy_content(self, revision_id=None):
2041
"""See InterRepository.copy_content()."""
2042
# weave specific optimised path:
2043
# TODO: jam 20070210 Internal, should be an assert, not translate
2044
revision_id = osutils.safe_revision_id(revision_id)
2046
self.target.set_make_working_trees(self.source.make_working_trees())
2047
except NotImplementedError:
2049
# FIXME do not peek!
2050
if self.source.control_files._transport.listable():
2051
pb = ui.ui_factory.nested_progress_bar()
2053
self.target.weave_store.copy_all_ids(
2054
self.source.weave_store,
2056
from_transaction=self.source.get_transaction(),
2057
to_transaction=self.target.get_transaction())
2058
pb.update('copying inventory', 0, 1)
2059
self.target.control_weaves.copy_multi(
2060
self.source.control_weaves, ['inventory'],
2061
from_transaction=self.source.get_transaction(),
2062
to_transaction=self.target.get_transaction())
2063
self.target._revision_store.text_store.copy_all_ids(
2064
self.source._revision_store.text_store,
2069
self.target.fetch(self.source, revision_id=revision_id)
2072
def fetch(self, revision_id=None, pb=None):
2073
"""See InterRepository.fetch()."""
2074
from bzrlib.fetch import GenericRepoFetcher
2075
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2076
self.source, self.source._format, self.target, self.target._format)
2077
# TODO: jam 20070210 This should be an assert, not a translate
2078
revision_id = osutils.safe_revision_id(revision_id)
2079
f = GenericRepoFetcher(to_repository=self.target,
2080
from_repository=self.source,
2081
last_revision=revision_id,
2083
return f.count_copied, f.failed_revisions
2086
def missing_revision_ids(self, revision_id=None):
2087
"""See InterRepository.missing_revision_ids()."""
2088
# we want all revisions to satisfy revision_id in source.
2089
# but we don't want to stat every file here and there.
2090
# we want then, all revisions other needs to satisfy revision_id
2091
# checked, but not those that we have locally.
2092
# so the first thing is to get a subset of the revisions to
2093
# satisfy revision_id in source, and then eliminate those that
2094
# we do already have.
2095
# this is slow on high latency connection to self, but as as this
2096
# disk format scales terribly for push anyway due to rewriting
2097
# inventory.weave, this is considered acceptable.
2099
if revision_id is not None:
2100
source_ids = self.source.get_ancestry(revision_id)
2101
assert source_ids[0] is None
2104
source_ids = self.source._all_possible_ids()
2105
source_ids_set = set(source_ids)
2106
# source_ids is the worst possible case we may need to pull.
2107
# now we want to filter source_ids against what we actually
2108
# have in target, but don't try to check for existence where we know
2109
# we do not have a revision as that would be pointless.
2110
target_ids = set(self.target._all_possible_ids())
2111
possibly_present_revisions = target_ids.intersection(source_ids_set)
2112
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2113
required_revisions = source_ids_set.difference(actually_present_revisions)
2114
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2115
if revision_id is not None:
2116
# we used get_ancestry to determine source_ids then we are assured all
2117
# revisions referenced are present as they are installed in topological order.
2118
# and the tip revision was validated by get_ancestry.
2119
return required_topo_revisions
2121
# if we just grabbed the possibly available ids, then
2122
# we only have an estimate of whats available and need to validate
2123
# that against the revision records.
2124
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2127
class InterKnitRepo(InterSameDataRepository):
2128
"""Optimised code paths between Knit based repositories."""
2131
def _get_repo_format_to_test(self):
2132
from bzrlib.repofmt import knitrepo
2133
return knitrepo.RepositoryFormatKnit1()
2136
def is_compatible(source, target):
2137
"""Be compatible with known Knit formats.
2139
We don't test for the stores being of specific types because that
2140
could lead to confusing results, and there is no need to be
2143
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
2145
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2146
isinstance(target._format, (RepositoryFormatKnit1)))
2147
except AttributeError:
2151
def fetch(self, revision_id=None, pb=None):
2152
"""See InterRepository.fetch()."""
2153
from bzrlib.fetch import KnitRepoFetcher
2154
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2155
self.source, self.source._format, self.target, self.target._format)
2156
# TODO: jam 20070210 This should be an assert, not a translate
2157
revision_id = osutils.safe_revision_id(revision_id)
2158
f = KnitRepoFetcher(to_repository=self.target,
2159
from_repository=self.source,
2160
last_revision=revision_id,
2162
return f.count_copied, f.failed_revisions
2165
def missing_revision_ids(self, revision_id=None):
2166
"""See InterRepository.missing_revision_ids()."""
2167
if revision_id is not None:
2168
source_ids = self.source.get_ancestry(revision_id)
2169
assert source_ids[0] is None
2172
source_ids = self.source.all_revision_ids()
2173
source_ids_set = set(source_ids)
2174
# source_ids is the worst possible case we may need to pull.
2175
# now we want to filter source_ids against what we actually
2176
# have in target, but don't try to check for existence where we know
2177
# we do not have a revision as that would be pointless.
2178
target_ids = set(self.target.all_revision_ids())
2179
possibly_present_revisions = target_ids.intersection(source_ids_set)
2180
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2181
required_revisions = source_ids_set.difference(actually_present_revisions)
2182
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2183
if revision_id is not None:
2184
# we used get_ancestry to determine source_ids then we are assured all
2185
# revisions referenced are present as they are installed in topological order.
2186
# and the tip revision was validated by get_ancestry.
2187
return required_topo_revisions
2189
# if we just grabbed the possibly available ids, then
2190
# we only have an estimate of whats available and need to validate
2191
# that against the revision records.
2192
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2195
class InterModel1and2(InterRepository):
2198
def _get_repo_format_to_test(self):
2202
def is_compatible(source, target):
2203
if not source.supports_rich_root() and target.supports_rich_root():
2209
def fetch(self, revision_id=None, pb=None):
2210
"""See InterRepository.fetch()."""
2211
from bzrlib.fetch import Model1toKnit2Fetcher
2212
# TODO: jam 20070210 This should be an assert, not a translate
2213
revision_id = osutils.safe_revision_id(revision_id)
2214
f = Model1toKnit2Fetcher(to_repository=self.target,
2215
from_repository=self.source,
2216
last_revision=revision_id,
2218
return f.count_copied, f.failed_revisions
2221
def copy_content(self, revision_id=None):
2222
"""Make a complete copy of the content in self into destination.
2224
This is a destructive operation! Do not use it on existing
2227
:param revision_id: Only copy the content needed to construct
2228
revision_id and its parents.
2231
self.target.set_make_working_trees(self.source.make_working_trees())
2232
except NotImplementedError:
2234
# TODO: jam 20070210 Internal, assert, don't translate
2235
revision_id = osutils.safe_revision_id(revision_id)
2236
# but don't bother fetching if we have the needed data now.
2237
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2238
self.target.has_revision(revision_id)):
2240
self.target.fetch(self.source, revision_id=revision_id)
2243
class InterKnit1and2(InterKnitRepo):
2246
def _get_repo_format_to_test(self):
2250
def is_compatible(source, target):
2251
"""Be compatible with Knit1 source and Knit3 target"""
2252
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2254
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
2255
RepositoryFormatKnit3
2256
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2257
isinstance(target._format, (RepositoryFormatKnit3)))
2258
except AttributeError:
2262
def fetch(self, revision_id=None, pb=None):
2263
"""See InterRepository.fetch()."""
2264
from bzrlib.fetch import Knit1to2Fetcher
2265
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2266
self.source, self.source._format, self.target,
2267
self.target._format)
2268
# TODO: jam 20070210 This should be an assert, not a translate
2269
revision_id = osutils.safe_revision_id(revision_id)
2270
f = Knit1to2Fetcher(to_repository=self.target,
2271
from_repository=self.source,
2272
last_revision=revision_id,
2274
return f.count_copied, f.failed_revisions
2277
class InterRemoteRepository(InterRepository):
2278
"""Code for converting between RemoteRepository objects.
2280
This just gets an non-remote repository from the RemoteRepository, and calls
2281
InterRepository.get again.
2284
def __init__(self, source, target):
2285
if isinstance(source, remote.RemoteRepository):
2286
source._ensure_real()
2287
real_source = source._real_repository
2289
real_source = source
2290
if isinstance(target, remote.RemoteRepository):
2291
target._ensure_real()
2292
real_target = target._real_repository
2294
real_target = target
2295
self.real_inter = InterRepository.get(real_source, real_target)
2298
def is_compatible(source, target):
2299
if isinstance(source, remote.RemoteRepository):
2301
if isinstance(target, remote.RemoteRepository):
2305
def copy_content(self, revision_id=None):
2306
self.real_inter.copy_content(revision_id=revision_id)
2308
def fetch(self, revision_id=None, pb=None):
2309
self.real_inter.fetch(revision_id=revision_id, pb=pb)
2312
def _get_repo_format_to_test(self):
2316
InterRepository.register_optimiser(InterSameDataRepository)
2317
InterRepository.register_optimiser(InterWeaveRepo)
2318
InterRepository.register_optimiser(InterKnitRepo)
2319
InterRepository.register_optimiser(InterModel1and2)
2320
InterRepository.register_optimiser(InterKnit1and2)
2321
InterRepository.register_optimiser(InterRemoteRepository)
2324
class CopyConverter(object):
2325
"""A repository conversion tool which just performs a copy of the content.
2327
This is slow but quite reliable.
2330
def __init__(self, target_format):
2331
"""Create a CopyConverter.
2333
:param target_format: The format the resulting repository should be.
2335
self.target_format = target_format
2337
def convert(self, repo, pb):
2338
"""Perform the conversion of to_convert, giving feedback via pb.
2340
:param to_convert: The disk object to convert.
2341
:param pb: a progress bar to use for progress information.
2346
# this is only useful with metadir layouts - separated repo content.
2347
# trigger an assertion if not such
2348
repo._format.get_format_string()
2349
self.repo_dir = repo.bzrdir
2350
self.step('Moving repository to repository.backup')
2351
self.repo_dir.transport.move('repository', 'repository.backup')
2352
backup_transport = self.repo_dir.transport.clone('repository.backup')
2353
repo._format.check_conversion_target(self.target_format)
2354
self.source_repo = repo._format.open(self.repo_dir,
2356
_override_transport=backup_transport)
2357
self.step('Creating new repository')
2358
converted = self.target_format.initialize(self.repo_dir,
2359
self.source_repo.is_shared())
2360
converted.lock_write()
2362
self.step('Copying content into repository.')
2363
self.source_repo.copy_content_into(converted)
2366
self.step('Deleting old repository content.')
2367
self.repo_dir.transport.delete_tree('repository.backup')
2368
self.pb.note('repository converted')
2370
def step(self, message):
2371
"""Update the pb by a step."""
2373
self.pb.update(message, self.count, self.total)
2385
def _unescaper(match, _map=_unescape_map):
2386
code = match.group(1)
2390
if not code.startswith('#'):
2392
return unichr(int(code[1:])).encode('utf8')
2398
def _unescape_xml(data):
2399
"""Unescape predefined XML entities in a string of data."""
2401
if _unescape_re is None:
2402
_unescape_re = re.compile('\&([^;]*);')
2403
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
