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,
45
from bzrlib.bundle import serializer
46
from bzrlib.revisiontree import RevisionTree
47
from bzrlib.store.versioned import VersionedFileStore
48
from bzrlib.store.text import TextStore
49
from bzrlib.testament import Testament
50
from bzrlib.util import bencode
53
from bzrlib.decorators import needs_read_lock, needs_write_lock
54
from bzrlib.inter import InterObject
55
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
56
from bzrlib.symbol_versioning import (
59
from bzrlib.trace import mutter, mutter_callsite, note, warning
62
# Old formats display a warning, but only once
63
_deprecation_warning_done = False
66
class CommitBuilder(object):
67
"""Provides an interface to build up a commit.
69
This allows describing a tree to be committed without needing to
70
know the internals of the format of the repository.
73
# all clients should supply tree roots.
74
record_root_entry = True
75
# the default CommitBuilder does not manage trees whose root is versioned.
76
_versioned_root = False
78
def __init__(self, repository, parents, config, timestamp=None,
79
timezone=None, committer=None, revprops=None,
81
"""Initiate a CommitBuilder.
83
:param repository: Repository to commit to.
84
:param parents: Revision ids of the parents of the new revision.
85
:param config: Configuration to use.
86
:param timestamp: Optional timestamp recorded for commit.
87
:param timezone: Optional timezone for timestamp.
88
:param committer: Optional committer to set for commit.
89
:param revprops: Optional dictionary of revision properties.
90
:param revision_id: Optional revision id.
95
self._committer = self._config.username()
97
self._committer = committer
99
self.new_inventory = Inventory(None)
100
self._new_revision_id = revision_id
101
self.parents = parents
102
self.repository = repository
105
if revprops is not None:
106
self._revprops.update(revprops)
108
if timestamp is None:
109
timestamp = time.time()
110
# Restrict resolution to 1ms
111
self._timestamp = round(timestamp, 3)
114
self._timezone = osutils.local_time_offset()
116
self._timezone = int(timezone)
118
self._generate_revision_if_needed()
119
self.__heads = graph.HeadsCache(repository.get_graph()).heads
121
def commit(self, message):
122
"""Make the actual commit.
124
:return: The revision id of the recorded revision.
126
rev = _mod_revision.Revision(
127
timestamp=self._timestamp,
128
timezone=self._timezone,
129
committer=self._committer,
131
inventory_sha1=self.inv_sha1,
132
revision_id=self._new_revision_id,
133
properties=self._revprops)
134
rev.parent_ids = self.parents
135
self.repository.add_revision(self._new_revision_id, rev,
136
self.new_inventory, self._config)
137
self.repository.commit_write_group()
138
return self._new_revision_id
141
"""Abort the commit that is being built.
143
self.repository.abort_write_group()
145
def revision_tree(self):
146
"""Return the tree that was just committed.
148
After calling commit() this can be called to get a RevisionTree
149
representing the newly committed tree. This is preferred to
150
calling Repository.revision_tree() because that may require
151
deserializing the inventory, while we already have a copy in
154
return RevisionTree(self.repository, self.new_inventory,
155
self._new_revision_id)
157
def finish_inventory(self):
158
"""Tell the builder that the inventory is finished."""
159
if self.new_inventory.root is None:
160
raise AssertionError('Root entry should be supplied to'
161
' record_entry_contents, as of bzr 0.10.',
162
DeprecationWarning, stacklevel=2)
163
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
164
self.new_inventory.revision_id = self._new_revision_id
165
self.inv_sha1 = self.repository.add_inventory(
166
self._new_revision_id,
171
def _gen_revision_id(self):
172
"""Return new revision-id."""
173
return generate_ids.gen_revision_id(self._config.username(),
176
def _generate_revision_if_needed(self):
177
"""Create a revision id if None was supplied.
179
If the repository can not support user-specified revision ids
180
they should override this function and raise CannotSetRevisionId
181
if _new_revision_id is not None.
183
:raises: CannotSetRevisionId
185
if self._new_revision_id is None:
186
self._new_revision_id = self._gen_revision_id()
187
self.random_revid = True
189
self.random_revid = False
191
def _heads(self, file_id, revision_ids):
192
"""Calculate the graph heads for revision_ids in the graph of file_id.
194
This can use either a per-file graph or a global revision graph as we
195
have an identity relationship between the two graphs.
197
return self.__heads(revision_ids)
199
def _check_root(self, ie, parent_invs, tree):
200
"""Helper for record_entry_contents.
202
:param ie: An entry being added.
203
:param parent_invs: The inventories of the parent revisions of the
205
:param tree: The tree that is being committed.
207
# In this revision format, root entries have no knit or weave When
208
# serializing out to disk and back in root.revision is always
210
ie.revision = self._new_revision_id
212
def _get_delta(self, ie, basis_inv, path):
213
"""Get a delta against the basis inventory for ie."""
214
if ie.file_id not in basis_inv:
216
return (None, path, ie.file_id, ie)
217
elif ie != basis_inv[ie.file_id]:
219
# TODO: avoid tis id2path call.
220
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
225
def record_entry_contents(self, ie, parent_invs, path, tree,
227
"""Record the content of ie from tree into the commit if needed.
229
Side effect: sets ie.revision when unchanged
231
:param ie: An inventory entry present in the commit.
232
:param parent_invs: The inventories of the parent revisions of the
234
:param path: The path the entry is at in the tree.
235
:param tree: The tree which contains this entry and should be used to
237
:param content_summary: Summary data from the tree about the paths
238
content - stat, length, exec, sha/link target. This is only
239
accessed when the entry has a revision of None - that is when it is
240
a candidate to commit.
241
:return: A tuple (change_delta, version_recorded). change_delta is
242
an inventory_delta change for this entry against the basis tree of
243
the commit, or None if no change occured against the basis tree.
244
version_recorded is True if a new version of the entry has been
245
recorded. For instance, committing a merge where a file was only
246
changed on the other side will return (delta, False).
248
if self.new_inventory.root is None:
249
if ie.parent_id is not None:
250
raise errors.RootMissing()
251
self._check_root(ie, parent_invs, tree)
252
if ie.revision is None:
253
kind = content_summary[0]
255
# ie is carried over from a prior commit
257
# XXX: repository specific check for nested tree support goes here - if
258
# the repo doesn't want nested trees we skip it ?
259
if (kind == 'tree-reference' and
260
not self.repository._format.supports_tree_reference):
261
# mismatch between commit builder logic and repository:
262
# this needs the entry creation pushed down into the builder.
263
raise NotImplementedError('Missing repository subtree support.')
264
self.new_inventory.add(ie)
266
# TODO: slow, take it out of the inner loop.
268
basis_inv = parent_invs[0]
270
basis_inv = Inventory(root_id=None)
272
# ie.revision is always None if the InventoryEntry is considered
273
# for committing. We may record the previous parents revision if the
274
# content is actually unchanged against a sole head.
275
if ie.revision is not None:
276
if not self._versioned_root and path == '':
277
# repositories that do not version the root set the root's
278
# revision to the new commit even when no change occurs, and
279
# this masks when a change may have occurred against the basis,
280
# so calculate if one happened.
281
if ie.file_id in basis_inv:
282
delta = (basis_inv.id2path(ie.file_id), path,
286
delta = (None, path, ie.file_id, ie)
289
# we don't need to commit this, because the caller already
290
# determined that an existing revision of this file is
292
return None, (ie.revision == self._new_revision_id)
293
# XXX: Friction: parent_candidates should return a list not a dict
294
# so that we don't have to walk the inventories again.
295
parent_candiate_entries = ie.parent_candidates(parent_invs)
296
head_set = self._heads(ie.file_id, parent_candiate_entries.keys())
298
for inv in parent_invs:
299
if ie.file_id in inv:
300
old_rev = inv[ie.file_id].revision
301
if old_rev in head_set:
302
heads.append(inv[ie.file_id].revision)
303
head_set.remove(inv[ie.file_id].revision)
306
# now we check to see if we need to write a new record to the
308
# We write a new entry unless there is one head to the ancestors, and
309
# the kind-derived content is unchanged.
311
# Cheapest check first: no ancestors, or more the one head in the
312
# ancestors, we write a new node.
316
# There is a single head, look it up for comparison
317
parent_entry = parent_candiate_entries[heads[0]]
318
# if the non-content specific data has changed, we'll be writing a
320
if (parent_entry.parent_id != ie.parent_id or
321
parent_entry.name != ie.name):
323
# now we need to do content specific checks:
325
# if the kind changed the content obviously has
326
if kind != parent_entry.kind:
329
if content_summary[2] is None:
330
raise ValueError("Files must not have executable = None")
332
if (# if the file length changed we have to store:
333
parent_entry.text_size != content_summary[1] or
334
# if the exec bit has changed we have to store:
335
parent_entry.executable != content_summary[2]):
337
elif parent_entry.text_sha1 == content_summary[3]:
338
# all meta and content is unchanged (using a hash cache
339
# hit to check the sha)
340
ie.revision = parent_entry.revision
341
ie.text_size = parent_entry.text_size
342
ie.text_sha1 = parent_entry.text_sha1
343
ie.executable = parent_entry.executable
344
return self._get_delta(ie, basis_inv, path), False
346
# Either there is only a hash change(no hash cache entry,
347
# or same size content change), or there is no change on
349
# Provide the parent's hash to the store layer, so that the
350
# content is unchanged we will not store a new node.
351
nostore_sha = parent_entry.text_sha1
353
# We want to record a new node regardless of the presence or
354
# absence of a content change in the file.
356
ie.executable = content_summary[2]
357
lines = tree.get_file(ie.file_id, path).readlines()
359
ie.text_sha1, ie.text_size = self._add_text_to_weave(
360
ie.file_id, lines, heads, nostore_sha)
361
except errors.ExistingContent:
362
# Turns out that the file content was unchanged, and we were
363
# only going to store a new node if it was changed. Carry over
365
ie.revision = parent_entry.revision
366
ie.text_size = parent_entry.text_size
367
ie.text_sha1 = parent_entry.text_sha1
368
ie.executable = parent_entry.executable
369
return self._get_delta(ie, basis_inv, path), False
370
elif kind == 'directory':
372
# all data is meta here, nothing specific to directory, so
374
ie.revision = parent_entry.revision
375
return self._get_delta(ie, basis_inv, path), False
377
self._add_text_to_weave(ie.file_id, lines, heads, None)
378
elif kind == 'symlink':
379
current_link_target = content_summary[3]
381
# symlink target is not generic metadata, check if it has
383
if current_link_target != parent_entry.symlink_target:
386
# unchanged, carry over.
387
ie.revision = parent_entry.revision
388
ie.symlink_target = parent_entry.symlink_target
389
return self._get_delta(ie, basis_inv, path), False
390
ie.symlink_target = current_link_target
392
self._add_text_to_weave(ie.file_id, lines, heads, None)
393
elif kind == 'tree-reference':
395
if content_summary[3] != parent_entry.reference_revision:
398
# unchanged, carry over.
399
ie.reference_revision = parent_entry.reference_revision
400
ie.revision = parent_entry.revision
401
return self._get_delta(ie, basis_inv, path), False
402
ie.reference_revision = content_summary[3]
404
self._add_text_to_weave(ie.file_id, lines, heads, None)
406
raise NotImplementedError('unknown kind')
407
ie.revision = self._new_revision_id
408
return self._get_delta(ie, basis_inv, path), True
410
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
411
versionedfile = self.repository.weave_store.get_weave_or_empty(
412
file_id, self.repository.get_transaction())
413
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
414
# than add_lines, and allows committing when a parent is ghosted for
416
# Note: as we read the content directly from the tree, we know its not
417
# been turned into unicode or badly split - but a broken tree
418
# implementation could give us bad output from readlines() so this is
419
# not a guarantee of safety. What would be better is always checking
420
# the content during test suite execution. RBC 20070912
421
return versionedfile.add_lines_with_ghosts(
422
self._new_revision_id, parents, new_lines,
423
nostore_sha=nostore_sha, random_id=self.random_revid,
424
check_content=False)[0:2]
427
class RootCommitBuilder(CommitBuilder):
428
"""This commitbuilder actually records the root id"""
430
# the root entry gets versioned properly by this builder.
431
_versioned_root = True
433
def _check_root(self, ie, parent_invs, tree):
434
"""Helper for record_entry_contents.
436
:param ie: An entry being added.
437
:param parent_invs: The inventories of the parent revisions of the
439
:param tree: The tree that is being committed.
443
######################################################################
446
class Repository(object):
447
"""Repository holding history for one or more branches.
449
The repository holds and retrieves historical information including
450
revisions and file history. It's normally accessed only by the Branch,
451
which views a particular line of development through that history.
453
The Repository builds on top of Stores and a Transport, which respectively
454
describe the disk data format and the way of accessing the (possibly
457
:ivar _transport: Transport for file access to repository, typically
458
pointing to .bzr/repository.
461
# What class to use for a CommitBuilder. Often its simpler to change this
462
# in a Repository class subclass rather than to override
463
# get_commit_builder.
464
_commit_builder_class = CommitBuilder
465
# The search regex used by xml based repositories to determine what things
466
# where changed in a single commit.
467
_file_ids_altered_regex = lazy_regex.lazy_compile(
468
r'file_id="(?P<file_id>[^"]+)"'
469
r'.* revision="(?P<revision_id>[^"]+)"'
472
def abort_write_group(self):
473
"""Commit the contents accrued within the current write group.
475
:seealso: start_write_group.
477
if self._write_group is not self.get_transaction():
478
# has an unlock or relock occured ?
479
raise errors.BzrError('mismatched lock context and write group.')
480
self._abort_write_group()
481
self._write_group = None
483
def _abort_write_group(self):
484
"""Template method for per-repository write group cleanup.
486
This is called during abort before the write group is considered to be
487
finished and should cleanup any internal state accrued during the write
488
group. There is no requirement that data handed to the repository be
489
*not* made available - this is not a rollback - but neither should any
490
attempt be made to ensure that data added is fully commited. Abort is
491
invoked when an error has occured so futher disk or network operations
492
may not be possible or may error and if possible should not be
496
def add_fallback_repository(self, repository):
497
"""Add a repository to use for looking up data not held locally.
499
:param repository: A repository.
501
if not self._format.supports_external_lookups:
502
raise errors.UnstackableRepositoryFormat(self._format, self.base)
503
if not self._add_fallback_repository_check(repository):
504
raise errors.IncompatibleRepositories(self, repository)
505
self._fallback_repositories.append(repository)
507
def _add_fallback_repository_check(self, repository):
508
"""Check that this repository can fallback to repository safely.
510
:param repository: A repository to fallback to.
511
:return: True if the repositories can stack ok.
513
return InterRepository._same_model(self, repository)
515
def add_inventory(self, revision_id, inv, parents):
516
"""Add the inventory inv to the repository as revision_id.
518
:param parents: The revision ids of the parents that revision_id
519
is known to have and are in the repository already.
521
:returns: The validator(which is a sha1 digest, though what is sha'd is
522
repository format specific) of the serialized inventory.
524
if not self.is_in_write_group():
525
raise AssertionError("%r not in write group" % (self,))
526
_mod_revision.check_not_reserved_id(revision_id)
527
if not (inv.revision_id is None or inv.revision_id == revision_id):
528
raise AssertionError(
529
"Mismatch between inventory revision"
530
" id and insertion revid (%r, %r)"
531
% (inv.revision_id, revision_id))
533
raise AssertionError()
534
inv_lines = self._serialise_inventory_to_lines(inv)
535
inv_vf = self.get_inventory_weave()
536
return self._inventory_add_lines(inv_vf, revision_id, parents,
537
inv_lines, check_content=False)
539
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
541
"""Store lines in inv_vf and return the sha1 of the inventory."""
543
for parent in parents:
545
final_parents.append(parent)
546
return inv_vf.add_lines(revision_id, final_parents, lines,
547
check_content=check_content)[0]
549
def add_revision(self, revision_id, rev, inv=None, config=None):
550
"""Add rev to the revision store as revision_id.
552
:param revision_id: the revision id to use.
553
:param rev: The revision object.
554
:param inv: The inventory for the revision. if None, it will be looked
555
up in the inventory storer
556
:param config: If None no digital signature will be created.
557
If supplied its signature_needed method will be used
558
to determine if a signature should be made.
560
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
562
_mod_revision.check_not_reserved_id(revision_id)
563
if config is not None and config.signature_needed():
565
inv = self.get_inventory(revision_id)
566
plaintext = Testament(rev, inv).as_short_text()
567
self.store_revision_signature(
568
gpg.GPGStrategy(config), plaintext, revision_id)
569
inventory_vf = self.get_inventory_weave()
570
if not revision_id in inventory_vf:
572
raise errors.WeaveRevisionNotPresent(revision_id,
575
# yes, this is not suitable for adding with ghosts.
576
rev.inventory_sha1 = self.add_inventory(revision_id, inv,
579
rev.inventory_sha1 = inventory_vf.get_sha1s([revision_id])[0]
580
self._revision_store.add_revision(rev, self.get_transaction())
582
def _add_revision_text(self, revision_id, text):
583
revision = self._revision_store._serializer.read_revision_from_string(
585
self._revision_store._add_revision(revision, StringIO(text),
586
self.get_transaction())
588
def all_revision_ids(self):
589
"""Returns a list of all the revision ids in the repository.
591
This is conceptually deprecated because code should generally work on
592
the graph reachable from a particular revision, and ignore any other
593
revisions that might be present. There is no direct replacement
596
if 'evil' in debug.debug_flags:
597
mutter_callsite(2, "all_revision_ids is linear with history.")
598
return self._all_revision_ids()
600
def _all_revision_ids(self):
601
"""Returns a list of all the revision ids in the repository.
603
These are in as much topological order as the underlying store can
606
raise NotImplementedError(self._all_revision_ids)
608
def break_lock(self):
609
"""Break a lock if one is present from another instance.
611
Uses the ui factory to ask for confirmation if the lock may be from
614
self.control_files.break_lock()
617
def _eliminate_revisions_not_present(self, revision_ids):
618
"""Check every revision id in revision_ids to see if we have it.
620
Returns a set of the present revisions.
623
graph = self.get_graph()
624
parent_map = graph.get_parent_map(revision_ids)
625
# The old API returned a list, should this actually be a set?
626
return parent_map.keys()
629
def create(a_bzrdir):
630
"""Construct the current default format repository in a_bzrdir."""
631
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
633
def __init__(self, _format, a_bzrdir, control_files,
634
_revision_store, control_store, text_store):
635
"""instantiate a Repository.
637
:param _format: The format of the repository on disk.
638
:param a_bzrdir: The BzrDir of the repository.
640
In the future we will have a single api for all stores for
641
getting file texts, inventories and revisions, then
642
this construct will accept instances of those things.
644
super(Repository, self).__init__()
645
self._format = _format
646
# the following are part of the public API for Repository:
647
self.bzrdir = a_bzrdir
648
self.control_files = control_files
649
self._transport = control_files._transport
650
self.base = self._transport.base
651
self._revision_store = _revision_store
652
# backwards compatibility
653
self.weave_store = text_store
655
self._reconcile_does_inventory_gc = True
656
self._reconcile_fixes_text_parents = False
657
self._reconcile_backsup_inventory = True
658
# not right yet - should be more semantically clear ?
660
self.control_store = control_store
661
self.control_weaves = control_store
662
# TODO: make sure to construct the right store classes, etc, depending
663
# on whether escaping is required.
664
self._warn_if_deprecated()
665
self._write_group = None
666
# Additional places to query for data.
667
self._fallback_repositories = []
670
return '%s(%r)' % (self.__class__.__name__,
673
def has_same_location(self, other):
674
"""Returns a boolean indicating if this repository is at the same
675
location as another repository.
677
This might return False even when two repository objects are accessing
678
the same physical repository via different URLs.
680
if self.__class__ is not other.__class__:
682
return (self._transport.base == other._transport.base)
684
def is_in_write_group(self):
685
"""Return True if there is an open write group.
687
:seealso: start_write_group.
689
return self._write_group is not None
692
return self.control_files.is_locked()
694
def is_write_locked(self):
695
"""Return True if this object is write locked."""
696
return self.is_locked() and self.control_files._lock_mode == 'w'
698
def lock_write(self, token=None):
699
"""Lock this repository for writing.
701
This causes caching within the repository obejct to start accumlating
702
data during reads, and allows a 'write_group' to be obtained. Write
703
groups must be used for actual data insertion.
705
:param token: if this is already locked, then lock_write will fail
706
unless the token matches the existing lock.
707
:returns: a token if this instance supports tokens, otherwise None.
708
:raises TokenLockingNotSupported: when a token is given but this
709
instance doesn't support using token locks.
710
:raises MismatchedToken: if the specified token doesn't match the token
711
of the existing lock.
712
:seealso: start_write_group.
714
A token should be passed in if you know that you have locked the object
715
some other way, and need to synchronise this object's state with that
718
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
720
result = self.control_files.lock_write(token=token)
725
self.control_files.lock_read()
728
def get_physical_lock_status(self):
729
return self.control_files.get_physical_lock_status()
731
def leave_lock_in_place(self):
732
"""Tell this repository not to release the physical lock when this
735
If lock_write doesn't return a token, then this method is not supported.
737
self.control_files.leave_in_place()
739
def dont_leave_lock_in_place(self):
740
"""Tell this repository to release the physical lock when this
741
object is unlocked, even if it didn't originally acquire it.
743
If lock_write doesn't return a token, then this method is not supported.
745
self.control_files.dont_leave_in_place()
748
def gather_stats(self, revid=None, committers=None):
749
"""Gather statistics from a revision id.
751
:param revid: The revision id to gather statistics from, if None, then
752
no revision specific statistics are gathered.
753
:param committers: Optional parameter controlling whether to grab
754
a count of committers from the revision specific statistics.
755
:return: A dictionary of statistics. Currently this contains:
756
committers: The number of committers if requested.
757
firstrev: A tuple with timestamp, timezone for the penultimate left
758
most ancestor of revid, if revid is not the NULL_REVISION.
759
latestrev: A tuple with timestamp, timezone for revid, if revid is
760
not the NULL_REVISION.
761
revisions: The total revision count in the repository.
762
size: An estimate disk size of the repository in bytes.
765
if revid and committers:
766
result['committers'] = 0
767
if revid and revid != _mod_revision.NULL_REVISION:
769
all_committers = set()
770
revisions = self.get_ancestry(revid)
771
# pop the leading None
773
first_revision = None
775
# ignore the revisions in the middle - just grab first and last
776
revisions = revisions[0], revisions[-1]
777
for revision in self.get_revisions(revisions):
778
if not first_revision:
779
first_revision = revision
781
all_committers.add(revision.committer)
782
last_revision = revision
784
result['committers'] = len(all_committers)
785
result['firstrev'] = (first_revision.timestamp,
786
first_revision.timezone)
787
result['latestrev'] = (last_revision.timestamp,
788
last_revision.timezone)
790
# now gather global repository information
791
if self.bzrdir.root_transport.listable():
792
c, t = self._revision_store.total_size(self.get_transaction())
793
result['revisions'] = c
797
def find_branches(self, using=False):
798
"""Find branches underneath this repository.
800
This will include branches inside other branches.
802
:param using: If True, list only branches using this repository.
804
if using and not self.is_shared():
806
return [self.bzrdir.open_branch()]
807
except errors.NotBranchError:
809
class Evaluator(object):
812
self.first_call = True
814
def __call__(self, bzrdir):
815
# On the first call, the parameter is always the bzrdir
816
# containing the current repo.
817
if not self.first_call:
819
repository = bzrdir.open_repository()
820
except errors.NoRepositoryPresent:
823
return False, (None, repository)
824
self.first_call = False
826
value = (bzrdir.open_branch(), None)
827
except errors.NotBranchError:
832
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
833
self.bzrdir.root_transport, evaluate=Evaluator()):
834
if branch is not None:
835
branches.append(branch)
836
if not using and repository is not None:
837
branches.extend(repository.find_branches())
840
def get_data_stream(self, revision_ids):
841
raise NotImplementedError(self.get_data_stream)
843
def get_data_stream_for_search(self, search_result):
844
"""Get a data stream that can be inserted to a repository.
846
:param search_result: A bzrlib.graph.SearchResult selecting the
848
:return: A data stream that can be inserted into a repository using
851
raise NotImplementedError(self.get_data_stream_for_search)
853
def insert_data_stream(self, stream):
854
"""XXX What does this really do?
856
Is it a substitute for fetch?
857
Should it manage its own write group ?
859
for item_key, bytes in stream:
860
if item_key[0] == 'file':
861
(file_id,) = item_key[1:]
862
knit = self.weave_store.get_weave_or_empty(
863
file_id, self.get_transaction())
864
elif item_key == ('inventory',):
865
knit = self.get_inventory_weave()
866
elif item_key == ('revisions',):
867
knit = self._revision_store.get_revision_file(
868
self.get_transaction())
869
elif item_key == ('signatures',):
870
knit = self._revision_store.get_signature_file(
871
self.get_transaction())
873
raise errors.RepositoryDataStreamError(
874
"Unrecognised data stream key '%s'" % (item_key,))
875
decoded_list = bencode.bdecode(bytes)
876
format = decoded_list.pop(0)
879
for version, options, parents, some_bytes in decoded_list:
880
data_list.append((version, options, len(some_bytes), parents))
881
knit_bytes += some_bytes
882
buffer = StringIO(knit_bytes)
883
def reader_func(count):
887
return buffer.read(count)
888
knit.insert_data_stream(
889
(format, data_list, reader_func))
892
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
893
"""Return the revision ids that other has that this does not.
895
These are returned in topological order.
897
revision_id: only return revision ids included by revision_id.
899
return InterRepository.get(other, self).search_missing_revision_ids(
900
revision_id, find_ghosts)
902
@deprecated_method(symbol_versioning.one_two)
904
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
905
"""Return the revision ids that other has that this does not.
907
These are returned in topological order.
909
revision_id: only return revision ids included by revision_id.
911
keys = self.search_missing_revision_ids(
912
other, revision_id, find_ghosts).get_keys()
915
parents = other.get_graph().get_parent_map(keys)
918
return tsort.topo_sort(parents)
922
"""Open the repository rooted at base.
924
For instance, if the repository is at URL/.bzr/repository,
925
Repository.open(URL) -> a Repository instance.
927
control = bzrdir.BzrDir.open(base)
928
return control.open_repository()
930
def copy_content_into(self, destination, revision_id=None):
931
"""Make a complete copy of the content in self into destination.
933
This is a destructive operation! Do not use it on existing
936
return InterRepository.get(self, destination).copy_content(revision_id)
938
def commit_write_group(self):
939
"""Commit the contents accrued within the current write group.
941
:seealso: start_write_group.
943
if self._write_group is not self.get_transaction():
944
# has an unlock or relock occured ?
945
raise errors.BzrError('mismatched lock context %r and '
947
(self.get_transaction(), self._write_group))
948
self._commit_write_group()
949
self._write_group = None
951
def _commit_write_group(self):
952
"""Template method for per-repository write group cleanup.
954
This is called before the write group is considered to be
955
finished and should ensure that all data handed to the repository
956
for writing during the write group is safely committed (to the
957
extent possible considering file system caching etc).
960
def fetch(self, source, revision_id=None, pb=None, find_ghosts=False):
961
"""Fetch the content required to construct revision_id from source.
963
If revision_id is None all content is copied.
964
:param find_ghosts: Find and copy revisions in the source that are
965
ghosts in the target (and not reachable directly by walking out to
966
the first-present revision in target from revision_id).
968
# fast path same-url fetch operations
969
if self.has_same_location(source):
970
# check that last_revision is in 'from' and then return a
972
if (revision_id is not None and
973
not _mod_revision.is_null(revision_id)):
974
self.get_revision(revision_id)
976
inter = InterRepository.get(source, self)
978
return inter.fetch(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts)
979
except NotImplementedError:
980
raise errors.IncompatibleRepositories(source, self)
982
def create_bundle(self, target, base, fileobj, format=None):
983
return serializer.write_bundle(self, target, base, fileobj, format)
985
def get_commit_builder(self, branch, parents, config, timestamp=None,
986
timezone=None, committer=None, revprops=None,
988
"""Obtain a CommitBuilder for this repository.
990
:param branch: Branch to commit to.
991
:param parents: Revision ids of the parents of the new revision.
992
:param config: Configuration to use.
993
:param timestamp: Optional timestamp recorded for commit.
994
:param timezone: Optional timezone for timestamp.
995
:param committer: Optional committer to set for commit.
996
:param revprops: Optional dictionary of revision properties.
997
:param revision_id: Optional revision id.
999
result = self._commit_builder_class(self, parents, config,
1000
timestamp, timezone, committer, revprops, revision_id)
1001
self.start_write_group()
1005
if (self.control_files._lock_count == 1 and
1006
self.control_files._lock_mode == 'w'):
1007
if self._write_group is not None:
1008
self.abort_write_group()
1009
self.control_files.unlock()
1010
raise errors.BzrError(
1011
'Must end write groups before releasing write locks.')
1012
self.control_files.unlock()
1015
def clone(self, a_bzrdir, revision_id=None):
1016
"""Clone this repository into a_bzrdir using the current format.
1018
Currently no check is made that the format of this repository and
1019
the bzrdir format are compatible. FIXME RBC 20060201.
1021
:return: The newly created destination repository.
1023
# TODO: deprecate after 0.16; cloning this with all its settings is
1024
# probably not very useful -- mbp 20070423
1025
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
1026
self.copy_content_into(dest_repo, revision_id)
1029
def start_write_group(self):
1030
"""Start a write group in the repository.
1032
Write groups are used by repositories which do not have a 1:1 mapping
1033
between file ids and backend store to manage the insertion of data from
1034
both fetch and commit operations.
1036
A write lock is required around the start_write_group/commit_write_group
1037
for the support of lock-requiring repository formats.
1039
One can only insert data into a repository inside a write group.
1043
if not self.is_write_locked():
1044
raise errors.NotWriteLocked(self)
1045
if self._write_group:
1046
raise errors.BzrError('already in a write group')
1047
self._start_write_group()
1048
# so we can detect unlock/relock - the write group is now entered.
1049
self._write_group = self.get_transaction()
1051
def _start_write_group(self):
1052
"""Template method for per-repository write group startup.
1054
This is called before the write group is considered to be
1059
def sprout(self, to_bzrdir, revision_id=None):
1060
"""Create a descendent repository for new development.
1062
Unlike clone, this does not copy the settings of the repository.
1064
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
1065
dest_repo.fetch(self, revision_id=revision_id)
1068
def _create_sprouting_repo(self, a_bzrdir, shared):
1069
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
1070
# use target default format.
1071
dest_repo = a_bzrdir.create_repository()
1073
# Most control formats need the repository to be specifically
1074
# created, but on some old all-in-one formats it's not needed
1076
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
1077
except errors.UninitializableFormat:
1078
dest_repo = a_bzrdir.open_repository()
1082
def has_revision(self, revision_id):
1083
"""True if this repository has a copy of the revision."""
1084
return revision_id in self.has_revisions((revision_id,))
1086
def has_revisions(self, revision_ids):
1087
"""Probe to find out the presence of multiple revisions.
1089
:param revision_ids: An iterable of revision_ids.
1090
:return: A set of the revision_ids that were present.
1092
raise NotImplementedError(self.has_revisions)
1094
return self._revision_store.has_revision_id(revision_id,
1095
self.get_transaction())
1098
def get_revision(self, revision_id):
1099
"""Return the Revision object for a named revision."""
1100
return self.get_revisions([revision_id])[0]
1103
def get_revision_reconcile(self, revision_id):
1104
"""'reconcile' helper routine that allows access to a revision always.
1106
This variant of get_revision does not cross check the weave graph
1107
against the revision one as get_revision does: but it should only
1108
be used by reconcile, or reconcile-alike commands that are correcting
1109
or testing the revision graph.
1111
return self._get_revisions([revision_id])[0]
1114
def get_revisions(self, revision_ids):
1115
"""Get many revisions at once."""
1116
return self._get_revisions(revision_ids)
1119
def _get_revisions(self, revision_ids):
1120
"""Core work logic to get many revisions without sanity checks."""
1121
for rev_id in revision_ids:
1122
if not rev_id or not isinstance(rev_id, basestring):
1123
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1124
revs = self._revision_store.get_revisions(revision_ids,
1125
self.get_transaction())
1129
def get_revision_xml(self, revision_id):
1130
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1131
# would have already do it.
1132
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1133
rev = self.get_revision(revision_id)
1134
rev_tmp = StringIO()
1135
# the current serializer..
1136
self._revision_store._serializer.write_revision(rev, rev_tmp)
1138
return rev_tmp.getvalue()
1140
def get_deltas_for_revisions(self, revisions):
1141
"""Produce a generator of revision deltas.
1143
Note that the input is a sequence of REVISIONS, not revision_ids.
1144
Trees will be held in memory until the generator exits.
1145
Each delta is relative to the revision's lefthand predecessor.
1147
required_trees = set()
1148
for revision in revisions:
1149
required_trees.add(revision.revision_id)
1150
required_trees.update(revision.parent_ids[:1])
1151
trees = dict((t.get_revision_id(), t) for
1152
t in self.revision_trees(required_trees))
1153
for revision in revisions:
1154
if not revision.parent_ids:
1155
old_tree = self.revision_tree(None)
1157
old_tree = trees[revision.parent_ids[0]]
1158
yield trees[revision.revision_id].changes_from(old_tree)
1161
def get_revision_delta(self, revision_id):
1162
"""Return the delta for one revision.
1164
The delta is relative to the left-hand predecessor of the
1167
r = self.get_revision(revision_id)
1168
return list(self.get_deltas_for_revisions([r]))[0]
1171
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1172
signature = gpg_strategy.sign(plaintext)
1173
self.add_signature_text(revision_id, signature)
1176
def add_signature_text(self, revision_id, signature):
1177
self._revision_store.add_revision_signature_text(revision_id,
1179
self.get_transaction())
1181
def find_text_key_references(self):
1182
"""Find the text key references within the repository.
1184
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1185
revision_ids. Each altered file-ids has the exact revision_ids that
1186
altered it listed explicitly.
1187
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1188
to whether they were referred to by the inventory of the
1189
revision_id that they contain. The inventory texts from all present
1190
revision ids are assessed to generate this report.
1192
revision_ids = self.all_revision_ids()
1193
w = self.get_inventory_weave()
1194
pb = ui.ui_factory.nested_progress_bar()
1196
return self._find_text_key_references_from_xml_inventory_lines(
1197
w.iter_lines_added_or_present_in_versions(revision_ids, pb=pb))
1201
def _find_text_key_references_from_xml_inventory_lines(self,
1203
"""Core routine for extracting references to texts from inventories.
1205
This performs the translation of xml lines to revision ids.
1207
:param line_iterator: An iterator of lines, origin_version_id
1208
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1209
to whether they were referred to by the inventory of the
1210
revision_id that they contain. Note that if that revision_id was
1211
not part of the line_iterator's output then False will be given -
1212
even though it may actually refer to that key.
1214
if not self._serializer.support_altered_by_hack:
1215
raise AssertionError(
1216
"_find_text_key_references_from_xml_inventory_lines only "
1217
"supported for branches which store inventory as unnested xml"
1218
", not on %r" % self)
1221
# this code needs to read every new line in every inventory for the
1222
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1223
# not present in one of those inventories is unnecessary but not
1224
# harmful because we are filtering by the revision id marker in the
1225
# inventory lines : we only select file ids altered in one of those
1226
# revisions. We don't need to see all lines in the inventory because
1227
# only those added in an inventory in rev X can contain a revision=X
1229
unescape_revid_cache = {}
1230
unescape_fileid_cache = {}
1232
# jam 20061218 In a big fetch, this handles hundreds of thousands
1233
# of lines, so it has had a lot of inlining and optimizing done.
1234
# Sorry that it is a little bit messy.
1235
# Move several functions to be local variables, since this is a long
1237
search = self._file_ids_altered_regex.search
1238
unescape = _unescape_xml
1239
setdefault = result.setdefault
1240
for line, version_id in line_iterator:
1241
match = search(line)
1244
# One call to match.group() returning multiple items is quite a
1245
# bit faster than 2 calls to match.group() each returning 1
1246
file_id, revision_id = match.group('file_id', 'revision_id')
1248
# Inlining the cache lookups helps a lot when you make 170,000
1249
# lines and 350k ids, versus 8.4 unique ids.
1250
# Using a cache helps in 2 ways:
1251
# 1) Avoids unnecessary decoding calls
1252
# 2) Re-uses cached strings, which helps in future set and
1254
# (2) is enough that removing encoding entirely along with
1255
# the cache (so we are using plain strings) results in no
1256
# performance improvement.
1258
revision_id = unescape_revid_cache[revision_id]
1260
unescaped = unescape(revision_id)
1261
unescape_revid_cache[revision_id] = unescaped
1262
revision_id = unescaped
1264
# Note that unconditionally unescaping means that we deserialise
1265
# every fileid, which for general 'pull' is not great, but we don't
1266
# really want to have some many fulltexts that this matters anyway.
1269
file_id = unescape_fileid_cache[file_id]
1271
unescaped = unescape(file_id)
1272
unescape_fileid_cache[file_id] = unescaped
1275
key = (file_id, revision_id)
1276
setdefault(key, False)
1277
if revision_id == version_id:
1281
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1283
"""Helper routine for fileids_altered_by_revision_ids.
1285
This performs the translation of xml lines to revision ids.
1287
:param line_iterator: An iterator of lines, origin_version_id
1288
:param revision_ids: The revision ids to filter for. This should be a
1289
set or other type which supports efficient __contains__ lookups, as
1290
the revision id from each parsed line will be looked up in the
1291
revision_ids filter.
1292
:return: a dictionary mapping altered file-ids to an iterable of
1293
revision_ids. Each altered file-ids has the exact revision_ids that
1294
altered it listed explicitly.
1297
setdefault = result.setdefault
1298
for file_id, revision_id in \
1299
self._find_text_key_references_from_xml_inventory_lines(
1300
line_iterator).iterkeys():
1301
# once data is all ensured-consistent; then this is
1302
# if revision_id == version_id
1303
if revision_id in revision_ids:
1304
setdefault(file_id, set()).add(revision_id)
1307
def fileids_altered_by_revision_ids(self, revision_ids, _inv_weave=None):
1308
"""Find the file ids and versions affected by revisions.
1310
:param revisions: an iterable containing revision ids.
1311
:param _inv_weave: The inventory weave from this repository or None.
1312
If None, the inventory weave will be opened automatically.
1313
:return: a dictionary mapping altered file-ids to an iterable of
1314
revision_ids. Each altered file-ids has the exact revision_ids that
1315
altered it listed explicitly.
1317
selected_revision_ids = set(revision_ids)
1318
w = _inv_weave or self.get_inventory_weave()
1319
pb = ui.ui_factory.nested_progress_bar()
1321
return self._find_file_ids_from_xml_inventory_lines(
1322
w.iter_lines_added_or_present_in_versions(
1323
selected_revision_ids, pb=pb),
1324
selected_revision_ids)
1328
def iter_files_bytes(self, desired_files):
1329
"""Iterate through file versions.
1331
Files will not necessarily be returned in the order they occur in
1332
desired_files. No specific order is guaranteed.
1334
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1335
value supplied by the caller as part of desired_files. It should
1336
uniquely identify the file version in the caller's context. (Examples:
1337
an index number or a TreeTransform trans_id.)
1339
bytes_iterator is an iterable of bytestrings for the file. The
1340
kind of iterable and length of the bytestrings are unspecified, but for
1341
this implementation, it is a list of lines produced by
1342
VersionedFile.get_lines().
1344
:param desired_files: a list of (file_id, revision_id, identifier)
1347
transaction = self.get_transaction()
1348
for file_id, revision_id, callable_data in desired_files:
1350
weave = self.weave_store.get_weave(file_id, transaction)
1351
except errors.NoSuchFile:
1352
raise errors.NoSuchIdInRepository(self, file_id)
1353
yield callable_data, weave.get_lines(revision_id)
1355
def _generate_text_key_index(self, text_key_references=None,
1357
"""Generate a new text key index for the repository.
1359
This is an expensive function that will take considerable time to run.
1361
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1362
list of parents, also text keys. When a given key has no parents,
1363
the parents list will be [NULL_REVISION].
1365
# All revisions, to find inventory parents.
1366
if ancestors is None:
1367
graph = self.get_graph()
1368
ancestors = graph.get_parent_map(self.all_revision_ids())
1369
if text_key_references is None:
1370
text_key_references = self.find_text_key_references()
1371
pb = ui.ui_factory.nested_progress_bar()
1373
return self._do_generate_text_key_index(ancestors,
1374
text_key_references, pb)
1378
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1379
"""Helper for _generate_text_key_index to avoid deep nesting."""
1380
revision_order = tsort.topo_sort(ancestors)
1381
invalid_keys = set()
1383
for revision_id in revision_order:
1384
revision_keys[revision_id] = set()
1385
text_count = len(text_key_references)
1386
# a cache of the text keys to allow reuse; costs a dict of all the
1387
# keys, but saves a 2-tuple for every child of a given key.
1389
for text_key, valid in text_key_references.iteritems():
1391
invalid_keys.add(text_key)
1393
revision_keys[text_key[1]].add(text_key)
1394
text_key_cache[text_key] = text_key
1395
del text_key_references
1397
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1398
NULL_REVISION = _mod_revision.NULL_REVISION
1399
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1400
# too small for large or very branchy trees. However, for 55K path
1401
# trees, it would be easy to use too much memory trivially. Ideally we
1402
# could gauge this by looking at available real memory etc, but this is
1403
# always a tricky proposition.
1404
inventory_cache = lru_cache.LRUCache(10)
1405
batch_size = 10 # should be ~150MB on a 55K path tree
1406
batch_count = len(revision_order) / batch_size + 1
1408
pb.update("Calculating text parents.", processed_texts, text_count)
1409
for offset in xrange(batch_count):
1410
to_query = revision_order[offset * batch_size:(offset + 1) *
1414
for rev_tree in self.revision_trees(to_query):
1415
revision_id = rev_tree.get_revision_id()
1416
parent_ids = ancestors[revision_id]
1417
for text_key in revision_keys[revision_id]:
1418
pb.update("Calculating text parents.", processed_texts)
1419
processed_texts += 1
1420
candidate_parents = []
1421
for parent_id in parent_ids:
1422
parent_text_key = (text_key[0], parent_id)
1424
check_parent = parent_text_key not in \
1425
revision_keys[parent_id]
1427
# the parent parent_id is a ghost:
1428
check_parent = False
1429
# truncate the derived graph against this ghost.
1430
parent_text_key = None
1432
# look at the parent commit details inventories to
1433
# determine possible candidates in the per file graph.
1436
inv = inventory_cache[parent_id]
1438
inv = self.revision_tree(parent_id).inventory
1439
inventory_cache[parent_id] = inv
1440
parent_entry = inv._byid.get(text_key[0], None)
1441
if parent_entry is not None:
1443
text_key[0], parent_entry.revision)
1445
parent_text_key = None
1446
if parent_text_key is not None:
1447
candidate_parents.append(
1448
text_key_cache[parent_text_key])
1449
parent_heads = text_graph.heads(candidate_parents)
1450
new_parents = list(parent_heads)
1451
new_parents.sort(key=lambda x:candidate_parents.index(x))
1452
if new_parents == []:
1453
new_parents = [NULL_REVISION]
1454
text_index[text_key] = new_parents
1456
for text_key in invalid_keys:
1457
text_index[text_key] = [NULL_REVISION]
1460
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1461
"""Get an iterable listing the keys of all the data introduced by a set
1464
The keys will be ordered so that the corresponding items can be safely
1465
fetched and inserted in that order.
1467
:returns: An iterable producing tuples of (knit-kind, file-id,
1468
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1469
'revisions'. file-id is None unless knit-kind is 'file'.
1471
# XXX: it's a bit weird to control the inventory weave caching in this
1472
# generator. Ideally the caching would be done in fetch.py I think. Or
1473
# maybe this generator should explicitly have the contract that it
1474
# should not be iterated until the previously yielded item has been
1476
inv_w = self.get_inventory_weave()
1478
# file ids that changed
1479
file_ids = self.fileids_altered_by_revision_ids(revision_ids, inv_w)
1481
num_file_ids = len(file_ids)
1482
for file_id, altered_versions in file_ids.iteritems():
1483
if _files_pb is not None:
1484
_files_pb.update("fetch texts", count, num_file_ids)
1486
yield ("file", file_id, altered_versions)
1487
# We're done with the files_pb. Note that it finished by the caller,
1488
# just as it was created by the caller.
1492
yield ("inventory", None, revision_ids)
1495
revisions_with_signatures = set()
1496
for rev_id in revision_ids:
1498
self.get_signature_text(rev_id)
1499
except errors.NoSuchRevision:
1503
revisions_with_signatures.add(rev_id)
1504
yield ("signatures", None, revisions_with_signatures)
1507
yield ("revisions", None, revision_ids)
1510
def get_inventory_weave(self):
1511
return self.control_weaves.get_weave('inventory',
1512
self.get_transaction())
1515
def get_inventory(self, revision_id):
1516
"""Get Inventory object by revision id."""
1517
return self.iter_inventories([revision_id]).next()
1519
def iter_inventories(self, revision_ids):
1520
"""Get many inventories by revision_ids.
1522
This will buffer some or all of the texts used in constructing the
1523
inventories in memory, but will only parse a single inventory at a
1526
:return: An iterator of inventories.
1528
if ((None in revision_ids)
1529
or (_mod_revision.NULL_REVISION in revision_ids)):
1530
raise ValueError('cannot get null revision inventory')
1531
return self._iter_inventories(revision_ids)
1533
def _iter_inventories(self, revision_ids):
1534
"""single-document based inventory iteration."""
1535
texts = self.get_inventory_weave().get_texts(revision_ids)
1536
for text, revision_id in zip(texts, revision_ids):
1537
yield self.deserialise_inventory(revision_id, text)
1539
def deserialise_inventory(self, revision_id, xml):
1540
"""Transform the xml into an inventory object.
1542
:param revision_id: The expected revision id of the inventory.
1543
:param xml: A serialised inventory.
1545
result = self._serializer.read_inventory_from_string(xml, revision_id)
1546
if result.revision_id != revision_id:
1547
raise AssertionError('revision id mismatch %s != %s' % (
1548
result.revision_id, revision_id))
1551
def serialise_inventory(self, inv):
1552
return self._serializer.write_inventory_to_string(inv)
1554
def _serialise_inventory_to_lines(self, inv):
1555
return self._serializer.write_inventory_to_lines(inv)
1557
def get_serializer_format(self):
1558
return self._serializer.format_num
1561
def get_inventory_xml(self, revision_id):
1562
"""Get inventory XML as a file object."""
1564
iw = self.get_inventory_weave()
1565
return iw.get_text(revision_id)
1567
raise errors.HistoryMissing(self, 'inventory', revision_id)
1570
def get_inventory_sha1(self, revision_id):
1571
"""Return the sha1 hash of the inventory entry
1573
return self.get_revision(revision_id).inventory_sha1
1575
def iter_reverse_revision_history(self, revision_id):
1576
"""Iterate backwards through revision ids in the lefthand history
1578
:param revision_id: The revision id to start with. All its lefthand
1579
ancestors will be traversed.
1581
graph = self.get_graph()
1582
next_id = revision_id
1584
if next_id in (None, _mod_revision.NULL_REVISION):
1587
# Note: The following line may raise KeyError in the event of
1588
# truncated history. We decided not to have a try:except:raise
1589
# RevisionNotPresent here until we see a use for it, because of the
1590
# cost in an inner loop that is by its very nature O(history).
1591
# Robert Collins 20080326
1592
parents = graph.get_parent_map([next_id])[next_id]
1593
if len(parents) == 0:
1596
next_id = parents[0]
1599
def get_revision_inventory(self, revision_id):
1600
"""Return inventory of a past revision."""
1601
# TODO: Unify this with get_inventory()
1602
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1603
# must be the same as its revision, so this is trivial.
1604
if revision_id is None:
1605
# This does not make sense: if there is no revision,
1606
# then it is the current tree inventory surely ?!
1607
# and thus get_root_id() is something that looks at the last
1608
# commit on the branch, and the get_root_id is an inventory check.
1609
raise NotImplementedError
1610
# return Inventory(self.get_root_id())
1612
return self.get_inventory(revision_id)
1615
def is_shared(self):
1616
"""Return True if this repository is flagged as a shared repository."""
1617
raise NotImplementedError(self.is_shared)
1620
def reconcile(self, other=None, thorough=False):
1621
"""Reconcile this repository."""
1622
from bzrlib.reconcile import RepoReconciler
1623
reconciler = RepoReconciler(self, thorough=thorough)
1624
reconciler.reconcile()
1627
def _refresh_data(self):
1628
"""Helper called from lock_* to ensure coherency with disk.
1630
The default implementation does nothing; it is however possible
1631
for repositories to maintain loaded indices across multiple locks
1632
by checking inside their implementation of this method to see
1633
whether their indices are still valid. This depends of course on
1634
the disk format being validatable in this manner.
1638
def revision_tree(self, revision_id):
1639
"""Return Tree for a revision on this branch.
1641
`revision_id` may be None for the empty tree revision.
1643
# TODO: refactor this to use an existing revision object
1644
# so we don't need to read it in twice.
1645
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1646
return RevisionTree(self, Inventory(root_id=None),
1647
_mod_revision.NULL_REVISION)
1649
inv = self.get_revision_inventory(revision_id)
1650
return RevisionTree(self, inv, revision_id)
1652
def revision_trees(self, revision_ids):
1653
"""Return Tree for a revision on this branch.
1655
`revision_id` may not be None or 'null:'"""
1656
inventories = self.iter_inventories(revision_ids)
1657
for inv in inventories:
1658
yield RevisionTree(self, inv, inv.revision_id)
1661
def get_ancestry(self, revision_id, topo_sorted=True):
1662
"""Return a list of revision-ids integrated by a revision.
1664
The first element of the list is always None, indicating the origin
1665
revision. This might change when we have history horizons, or
1666
perhaps we should have a new API.
1668
This is topologically sorted.
1670
if _mod_revision.is_null(revision_id):
1672
if not self.has_revision(revision_id):
1673
raise errors.NoSuchRevision(self, revision_id)
1674
w = self.get_inventory_weave()
1675
candidates = w.get_ancestry(revision_id, topo_sorted)
1676
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1679
"""Compress the data within the repository.
1681
This operation only makes sense for some repository types. For other
1682
types it should be a no-op that just returns.
1684
This stub method does not require a lock, but subclasses should use
1685
@needs_write_lock as this is a long running call its reasonable to
1686
implicitly lock for the user.
1690
def print_file(self, file, revision_id):
1691
"""Print `file` to stdout.
1693
FIXME RBC 20060125 as John Meinel points out this is a bad api
1694
- it writes to stdout, it assumes that that is valid etc. Fix
1695
by creating a new more flexible convenience function.
1697
tree = self.revision_tree(revision_id)
1698
# use inventory as it was in that revision
1699
file_id = tree.inventory.path2id(file)
1701
# TODO: jam 20060427 Write a test for this code path
1702
# it had a bug in it, and was raising the wrong
1704
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1705
tree.print_file(file_id)
1707
def get_transaction(self):
1708
return self.control_files.get_transaction()
1710
@deprecated_method(symbol_versioning.one_one)
1711
def get_parents(self, revision_ids):
1712
"""See StackedParentsProvider.get_parents"""
1713
parent_map = self.get_parent_map(revision_ids)
1714
return [parent_map.get(r, None) for r in revision_ids]
1716
def get_parent_map(self, keys):
1717
"""See graph._StackedParentsProvider.get_parent_map"""
1719
for revision_id in keys:
1720
if revision_id is None:
1721
raise ValueError('get_parent_map(None) is not valid')
1722
if revision_id == _mod_revision.NULL_REVISION:
1723
parent_map[revision_id] = ()
1726
parent_id_list = self.get_revision(revision_id).parent_ids
1727
except errors.NoSuchRevision:
1730
if len(parent_id_list) == 0:
1731
parent_ids = (_mod_revision.NULL_REVISION,)
1733
parent_ids = tuple(parent_id_list)
1734
parent_map[revision_id] = parent_ids
1737
def _make_parents_provider(self):
1740
def get_graph(self, other_repository=None):
1741
"""Return the graph walker for this repository format"""
1742
parents_provider = self._make_parents_provider()
1743
if (other_repository is not None and
1744
not self.has_same_location(other_repository)):
1745
parents_provider = graph._StackedParentsProvider(
1746
[parents_provider, other_repository._make_parents_provider()])
1747
return graph.Graph(parents_provider)
1749
def _get_versioned_file_checker(self):
1750
"""Return an object suitable for checking versioned files."""
1751
return _VersionedFileChecker(self)
1753
def revision_ids_to_search_result(self, result_set):
1754
"""Convert a set of revision ids to a graph SearchResult."""
1755
result_parents = set()
1756
for parents in self.get_graph().get_parent_map(
1757
result_set).itervalues():
1758
result_parents.update(parents)
1759
included_keys = result_set.intersection(result_parents)
1760
start_keys = result_set.difference(included_keys)
1761
exclude_keys = result_parents.difference(result_set)
1762
result = graph.SearchResult(start_keys, exclude_keys,
1763
len(result_set), result_set)
1767
def set_make_working_trees(self, new_value):
1768
"""Set the policy flag for making working trees when creating branches.
1770
This only applies to branches that use this repository.
1772
The default is 'True'.
1773
:param new_value: True to restore the default, False to disable making
1776
raise NotImplementedError(self.set_make_working_trees)
1778
def make_working_trees(self):
1779
"""Returns the policy for making working trees on new branches."""
1780
raise NotImplementedError(self.make_working_trees)
1783
def sign_revision(self, revision_id, gpg_strategy):
1784
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1785
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1788
def has_signature_for_revision_id(self, revision_id):
1789
"""Query for a revision signature for revision_id in the repository."""
1790
return self._revision_store.has_signature(revision_id,
1791
self.get_transaction())
1794
def get_signature_text(self, revision_id):
1795
"""Return the text for a signature."""
1796
return self._revision_store.get_signature_text(revision_id,
1797
self.get_transaction())
1800
def check(self, revision_ids=None):
1801
"""Check consistency of all history of given revision_ids.
1803
Different repository implementations should override _check().
1805
:param revision_ids: A non-empty list of revision_ids whose ancestry
1806
will be checked. Typically the last revision_id of a branch.
1808
return self._check(revision_ids)
1810
def _check(self, revision_ids):
1811
result = check.Check(self)
1815
def _warn_if_deprecated(self):
1816
global _deprecation_warning_done
1817
if _deprecation_warning_done:
1819
_deprecation_warning_done = True
1820
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1821
% (self._format, self.bzrdir.transport.base))
1823
def supports_rich_root(self):
1824
return self._format.rich_root_data
1826
def _check_ascii_revisionid(self, revision_id, method):
1827
"""Private helper for ascii-only repositories."""
1828
# weave repositories refuse to store revisionids that are non-ascii.
1829
if revision_id is not None:
1830
# weaves require ascii revision ids.
1831
if isinstance(revision_id, unicode):
1833
revision_id.encode('ascii')
1834
except UnicodeEncodeError:
1835
raise errors.NonAsciiRevisionId(method, self)
1838
revision_id.decode('ascii')
1839
except UnicodeDecodeError:
1840
raise errors.NonAsciiRevisionId(method, self)
1842
def revision_graph_can_have_wrong_parents(self):
1843
"""Is it possible for this repository to have a revision graph with
1846
If True, then this repository must also implement
1847
_find_inconsistent_revision_parents so that check and reconcile can
1848
check for inconsistencies before proceeding with other checks that may
1849
depend on the revision index being consistent.
1851
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1854
# remove these delegates a while after bzr 0.15
1855
def __make_delegated(name, from_module):
1856
def _deprecated_repository_forwarder():
1857
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1858
% (name, from_module),
1861
m = __import__(from_module, globals(), locals(), [name])
1863
return getattr(m, name)
1864
except AttributeError:
1865
raise AttributeError('module %s has no name %s'
1867
globals()[name] = _deprecated_repository_forwarder
1870
'AllInOneRepository',
1871
'WeaveMetaDirRepository',
1872
'PreSplitOutRepositoryFormat',
1873
'RepositoryFormat4',
1874
'RepositoryFormat5',
1875
'RepositoryFormat6',
1876
'RepositoryFormat7',
1878
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1882
'RepositoryFormatKnit',
1883
'RepositoryFormatKnit1',
1885
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1888
def install_revision(repository, rev, revision_tree):
1889
"""Install all revision data into a repository."""
1890
install_revisions(repository, [(rev, revision_tree, None)])
1893
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1894
"""Install all revision data into a repository.
1896
Accepts an iterable of revision, tree, signature tuples. The signature
1899
repository.start_write_group()
1901
for n, (revision, revision_tree, signature) in enumerate(iterable):
1902
_install_revision(repository, revision, revision_tree, signature)
1904
pb.update('Transferring revisions', n + 1, num_revisions)
1906
repository.abort_write_group()
1909
repository.commit_write_group()
1912
def _install_revision(repository, rev, revision_tree, signature):
1913
"""Install all revision data into a repository."""
1914
present_parents = []
1916
for p_id in rev.parent_ids:
1917
if repository.has_revision(p_id):
1918
present_parents.append(p_id)
1919
parent_trees[p_id] = repository.revision_tree(p_id)
1921
parent_trees[p_id] = repository.revision_tree(None)
1923
inv = revision_tree.inventory
1924
entries = inv.iter_entries()
1925
# backwards compatibility hack: skip the root id.
1926
if not repository.supports_rich_root():
1927
path, root = entries.next()
1928
if root.revision != rev.revision_id:
1929
raise errors.IncompatibleRevision(repr(repository))
1930
# Add the texts that are not already present
1931
for path, ie in entries:
1932
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1933
repository.get_transaction())
1934
if ie.revision not in w:
1936
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1937
# with InventoryEntry.find_previous_heads(). if it is, then there
1938
# is a latent bug here where the parents may have ancestors of each
1940
for revision, tree in parent_trees.iteritems():
1941
if ie.file_id not in tree:
1943
parent_id = tree.inventory[ie.file_id].revision
1944
if parent_id in text_parents:
1946
text_parents.append(parent_id)
1948
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1949
repository.get_transaction())
1950
lines = revision_tree.get_file(ie.file_id).readlines()
1951
vfile.add_lines(rev.revision_id, text_parents, lines)
1953
# install the inventory
1954
repository.add_inventory(rev.revision_id, inv, present_parents)
1955
except errors.RevisionAlreadyPresent:
1957
if signature is not None:
1958
repository.add_signature_text(rev.revision_id, signature)
1959
repository.add_revision(rev.revision_id, rev, inv)
1962
class MetaDirRepository(Repository):
1963
"""Repositories in the new meta-dir layout.
1965
:ivar _transport: Transport for access to repository control files,
1966
typically pointing to .bzr/repository.
1969
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1970
super(MetaDirRepository, self).__init__(_format,
1976
self._transport = control_files._transport
1979
def is_shared(self):
1980
"""Return True if this repository is flagged as a shared repository."""
1981
return self._transport.has('shared-storage')
1984
def set_make_working_trees(self, new_value):
1985
"""Set the policy flag for making working trees when creating branches.
1987
This only applies to branches that use this repository.
1989
The default is 'True'.
1990
:param new_value: True to restore the default, False to disable making
1995
self._transport.delete('no-working-trees')
1996
except errors.NoSuchFile:
1999
self._transport.put_bytes('no-working-trees', '',
2000
mode=self.bzrdir._get_file_mode())
2002
def make_working_trees(self):
2003
"""Returns the policy for making working trees on new branches."""
2004
return not self._transport.has('no-working-trees')
2007
class MetaDirVersionedFileRepository(MetaDirRepository):
2008
"""Repositories in a meta-dir, that work via versioned file objects."""
2010
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
2011
super(MetaDirVersionedFileRepository, self).__init__(_format, a_bzrdir,
2012
control_files, _revision_store, control_store, text_store)
2013
_revision_store.get_scope = self.get_transaction
2014
control_store.get_scope = self.get_transaction
2015
text_store.get_scope = self.get_transaction
2018
class RepositoryFormatRegistry(registry.Registry):
2019
"""Registry of RepositoryFormats."""
2021
def get(self, format_string):
2022
r = registry.Registry.get(self, format_string)
2028
format_registry = RepositoryFormatRegistry()
2029
"""Registry of formats, indexed by their identifying format string.
2031
This can contain either format instances themselves, or classes/factories that
2032
can be called to obtain one.
2036
#####################################################################
2037
# Repository Formats
2039
class RepositoryFormat(object):
2040
"""A repository format.
2042
Formats provide three things:
2043
* An initialization routine to construct repository data on disk.
2044
* a format string which is used when the BzrDir supports versioned
2046
* an open routine which returns a Repository instance.
2048
There is one and only one Format subclass for each on-disk format. But
2049
there can be one Repository subclass that is used for several different
2050
formats. The _format attribute on a Repository instance can be used to
2051
determine the disk format.
2053
Formats are placed in an dict by their format string for reference
2054
during opening. These should be subclasses of RepositoryFormat
2057
Once a format is deprecated, just deprecate the initialize and open
2058
methods on the format class. Do not deprecate the object, as the
2059
object will be created every system load.
2061
Common instance attributes:
2062
_matchingbzrdir - the bzrdir format that the repository format was
2063
originally written to work with. This can be used if manually
2064
constructing a bzrdir and repository, or more commonly for test suite
2068
# Set to True or False in derived classes. True indicates that the format
2069
# supports ghosts gracefully.
2070
supports_ghosts = None
2071
# Can this repository be given external locations to lookup additional
2072
# data. Set to True or False in derived classes.
2073
supports_external_lookups = None
2076
return "<%s>" % self.__class__.__name__
2078
def __eq__(self, other):
2079
# format objects are generally stateless
2080
return isinstance(other, self.__class__)
2082
def __ne__(self, other):
2083
return not self == other
2086
def find_format(klass, a_bzrdir):
2087
"""Return the format for the repository object in a_bzrdir.
2089
This is used by bzr native formats that have a "format" file in
2090
the repository. Other methods may be used by different types of
2094
transport = a_bzrdir.get_repository_transport(None)
2095
format_string = transport.get("format").read()
2096
return format_registry.get(format_string)
2097
except errors.NoSuchFile:
2098
raise errors.NoRepositoryPresent(a_bzrdir)
2100
raise errors.UnknownFormatError(format=format_string,
2104
def register_format(klass, format):
2105
format_registry.register(format.get_format_string(), format)
2108
def unregister_format(klass, format):
2109
format_registry.remove(format.get_format_string())
2112
def get_default_format(klass):
2113
"""Return the current default format."""
2114
from bzrlib import bzrdir
2115
return bzrdir.format_registry.make_bzrdir('default').repository_format
2117
def _get_control_store(self, repo_transport, control_files):
2118
"""Return the control store for this repository."""
2119
raise NotImplementedError(self._get_control_store)
2121
def get_format_string(self):
2122
"""Return the ASCII format string that identifies this format.
2124
Note that in pre format ?? repositories the format string is
2125
not permitted nor written to disk.
2127
raise NotImplementedError(self.get_format_string)
2129
def get_format_description(self):
2130
"""Return the short description for this format."""
2131
raise NotImplementedError(self.get_format_description)
2133
def _get_revision_store(self, repo_transport, control_files):
2134
"""Return the revision store object for this a_bzrdir."""
2135
raise NotImplementedError(self._get_revision_store)
2137
def _get_text_rev_store(self,
2144
"""Common logic for getting a revision store for a repository.
2146
see self._get_revision_store for the subclass-overridable method to
2147
get the store for a repository.
2149
from bzrlib.store.revision.text import TextRevisionStore
2150
dir_mode = control_files._dir_mode
2151
file_mode = control_files._file_mode
2152
text_store = TextStore(transport.clone(name),
2154
compressed=compressed,
2156
file_mode=file_mode)
2157
_revision_store = TextRevisionStore(text_store, serializer)
2158
return _revision_store
2160
# TODO: this shouldn't be in the base class, it's specific to things that
2161
# use weaves or knits -- mbp 20070207
2162
def _get_versioned_file_store(self,
2167
versionedfile_class=None,
2168
versionedfile_kwargs={},
2170
if versionedfile_class is None:
2171
versionedfile_class = self._versionedfile_class
2172
weave_transport = control_files._transport.clone(name)
2173
dir_mode = control_files._dir_mode
2174
file_mode = control_files._file_mode
2175
return VersionedFileStore(weave_transport, prefixed=prefixed,
2177
file_mode=file_mode,
2178
versionedfile_class=versionedfile_class,
2179
versionedfile_kwargs=versionedfile_kwargs,
2182
def initialize(self, a_bzrdir, shared=False):
2183
"""Initialize a repository of this format in a_bzrdir.
2185
:param a_bzrdir: The bzrdir to put the new repository in it.
2186
:param shared: The repository should be initialized as a sharable one.
2187
:returns: The new repository object.
2189
This may raise UninitializableFormat if shared repository are not
2190
compatible the a_bzrdir.
2192
raise NotImplementedError(self.initialize)
2194
def is_supported(self):
2195
"""Is this format supported?
2197
Supported formats must be initializable and openable.
2198
Unsupported formats may not support initialization or committing or
2199
some other features depending on the reason for not being supported.
2203
def check_conversion_target(self, target_format):
2204
raise NotImplementedError(self.check_conversion_target)
2206
def open(self, a_bzrdir, _found=False):
2207
"""Return an instance of this format for the bzrdir a_bzrdir.
2209
_found is a private parameter, do not use it.
2211
raise NotImplementedError(self.open)
2214
class MetaDirRepositoryFormat(RepositoryFormat):
2215
"""Common base class for the new repositories using the metadir layout."""
2217
rich_root_data = False
2218
supports_tree_reference = False
2219
supports_external_lookups = False
2220
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
2223
super(MetaDirRepositoryFormat, self).__init__()
2225
def _create_control_files(self, a_bzrdir):
2226
"""Create the required files and the initial control_files object."""
2227
# FIXME: RBC 20060125 don't peek under the covers
2228
# NB: no need to escape relative paths that are url safe.
2229
repository_transport = a_bzrdir.get_repository_transport(self)
2230
control_files = lockable_files.LockableFiles(repository_transport,
2231
'lock', lockdir.LockDir)
2232
control_files.create_lock()
2233
return control_files
2235
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
2236
"""Upload the initial blank content."""
2237
control_files = self._create_control_files(a_bzrdir)
2238
control_files.lock_write()
2239
transport = control_files._transport
2241
utf8_files += [('shared-storage', '')]
2243
transport.mkdir_multi(dirs, mode=a_bzrdir._get_dir_mode())
2244
for (filename, content_stream) in files:
2245
transport.put_file(filename, content_stream,
2246
mode=a_bzrdir._get_file_mode())
2247
for (filename, content_bytes) in utf8_files:
2248
transport.put_bytes_non_atomic(filename, content_bytes,
2249
mode=a_bzrdir._get_file_mode())
2251
control_files.unlock()
2254
# formats which have no format string are not discoverable
2255
# and not independently creatable, so are not registered. They're
2256
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
2257
# needed, it's constructed directly by the BzrDir. Non-native formats where
2258
# the repository is not separately opened are similar.
2260
format_registry.register_lazy(
2261
'Bazaar-NG Repository format 7',
2262
'bzrlib.repofmt.weaverepo',
2266
format_registry.register_lazy(
2267
'Bazaar-NG Knit Repository Format 1',
2268
'bzrlib.repofmt.knitrepo',
2269
'RepositoryFormatKnit1',
2272
format_registry.register_lazy(
2273
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
2274
'bzrlib.repofmt.knitrepo',
2275
'RepositoryFormatKnit3',
2278
format_registry.register_lazy(
2279
'Bazaar Knit Repository Format 4 (bzr 1.0)\n',
2280
'bzrlib.repofmt.knitrepo',
2281
'RepositoryFormatKnit4',
2284
# Pack-based formats. There is one format for pre-subtrees, and one for
2285
# post-subtrees to allow ease of testing.
2286
# NOTE: These are experimental in 0.92. Stable in 1.0 and above
2287
format_registry.register_lazy(
2288
'Bazaar pack repository format 1 (needs bzr 0.92)\n',
2289
'bzrlib.repofmt.pack_repo',
2290
'RepositoryFormatKnitPack1',
2292
format_registry.register_lazy(
2293
'Bazaar pack repository format 1 with subtree support (needs bzr 0.92)\n',
2294
'bzrlib.repofmt.pack_repo',
2295
'RepositoryFormatKnitPack3',
2297
format_registry.register_lazy(
2298
'Bazaar pack repository format 1 with rich root (needs bzr 1.0)\n',
2299
'bzrlib.repofmt.pack_repo',
2300
'RepositoryFormatKnitPack4',
2302
# Development formats.
2304
# development 0 - stub to introduce development versioning scheme.
2305
format_registry.register_lazy(
2306
"Bazaar development format 0 (needs bzr.dev from before 1.3)\n",
2307
'bzrlib.repofmt.pack_repo',
2308
'RepositoryFormatPackDevelopment0',
2310
format_registry.register_lazy(
2311
("Bazaar development format 0 with subtree support "
2312
"(needs bzr.dev from before 1.3)\n"),
2313
'bzrlib.repofmt.pack_repo',
2314
'RepositoryFormatPackDevelopment0Subtree',
2316
format_registry.register_lazy(
2317
"Bazaar development format 1 (needs bzr.dev from before 1.6)\n",
2318
'bzrlib.repofmt.pack_repo',
2319
'RepositoryFormatPackDevelopment1',
2321
format_registry.register_lazy(
2322
("Bazaar development format 1 with subtree support "
2323
"(needs bzr.dev from before 1.6)\n"),
2324
'bzrlib.repofmt.pack_repo',
2325
'RepositoryFormatPackDevelopment1Subtree',
2327
# 1.3->1.4 go below here
2330
class InterRepository(InterObject):
2331
"""This class represents operations taking place between two repositories.
2333
Its instances have methods like copy_content and fetch, and contain
2334
references to the source and target repositories these operations can be
2337
Often we will provide convenience methods on 'repository' which carry out
2338
operations with another repository - they will always forward to
2339
InterRepository.get(other).method_name(parameters).
2343
"""The available optimised InterRepository types."""
2345
def copy_content(self, revision_id=None):
2346
raise NotImplementedError(self.copy_content)
2348
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2349
"""Fetch the content required to construct revision_id.
2351
The content is copied from self.source to self.target.
2353
:param revision_id: if None all content is copied, if NULL_REVISION no
2355
:param pb: optional progress bar to use for progress reports. If not
2356
provided a default one will be created.
2358
Returns the copied revision count and the failed revisions in a tuple:
2361
raise NotImplementedError(self.fetch)
2363
def _walk_to_common_revisions(self, revision_ids):
2364
"""Walk out from revision_ids in source to revisions target has.
2366
:param revision_ids: The start point for the search.
2367
:return: A set of revision ids.
2369
target_graph = self.target.get_graph()
2370
revision_ids = frozenset(revision_ids)
2371
if set(target_graph.get_parent_map(revision_ids)) == revision_ids:
2372
return graph.SearchResult(revision_ids, set(), 0, set())
2373
missing_revs = set()
2374
source_graph = self.source.get_graph()
2375
# ensure we don't pay silly lookup costs.
2376
searcher = source_graph._make_breadth_first_searcher(revision_ids)
2377
null_set = frozenset([_mod_revision.NULL_REVISION])
2380
next_revs, ghosts = searcher.next_with_ghosts()
2381
except StopIteration:
2383
if revision_ids.intersection(ghosts):
2384
absent_ids = set(revision_ids.intersection(ghosts))
2385
# If all absent_ids are present in target, no error is needed.
2386
absent_ids.difference_update(
2387
set(target_graph.get_parent_map(absent_ids)))
2389
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2390
# we don't care about other ghosts as we can't fetch them and
2391
# haven't been asked to.
2392
next_revs = set(next_revs)
2393
# we always have NULL_REVISION present.
2394
have_revs = set(target_graph.get_parent_map(next_revs)).union(null_set)
2395
missing_revs.update(next_revs - have_revs)
2396
searcher.stop_searching_any(have_revs)
2397
return searcher.get_result()
2399
@deprecated_method(symbol_versioning.one_two)
2401
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2402
"""Return the revision ids that source has that target does not.
2404
These are returned in topological order.
2406
:param revision_id: only return revision ids included by this
2408
:param find_ghosts: If True find missing revisions in deep history
2409
rather than just finding the surface difference.
2411
return list(self.search_missing_revision_ids(
2412
revision_id, find_ghosts).get_keys())
2415
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2416
"""Return the revision ids that source has that target does not.
2418
:param revision_id: only return revision ids included by this
2420
:param find_ghosts: If True find missing revisions in deep history
2421
rather than just finding the surface difference.
2422
:return: A bzrlib.graph.SearchResult.
2424
# stop searching at found target revisions.
2425
if not find_ghosts and revision_id is not None:
2426
return self._walk_to_common_revisions([revision_id])
2427
# generic, possibly worst case, slow code path.
2428
target_ids = set(self.target.all_revision_ids())
2429
if revision_id is not None:
2430
source_ids = self.source.get_ancestry(revision_id)
2431
if source_ids[0] is not None:
2432
raise AssertionError()
2435
source_ids = self.source.all_revision_ids()
2436
result_set = set(source_ids).difference(target_ids)
2437
return self.source.revision_ids_to_search_result(result_set)
2440
def _same_model(source, target):
2441
"""True if source and target have the same data representation."""
2442
if source.supports_rich_root() != target.supports_rich_root():
2444
if source._serializer != target._serializer:
2449
class InterSameDataRepository(InterRepository):
2450
"""Code for converting between repositories that represent the same data.
2452
Data format and model must match for this to work.
2456
def _get_repo_format_to_test(self):
2457
"""Repository format for testing with.
2459
InterSameData can pull from subtree to subtree and from non-subtree to
2460
non-subtree, so we test this with the richest repository format.
2462
from bzrlib.repofmt import knitrepo
2463
return knitrepo.RepositoryFormatKnit3()
2466
def is_compatible(source, target):
2467
return InterRepository._same_model(source, target)
2470
def copy_content(self, revision_id=None):
2471
"""Make a complete copy of the content in self into destination.
2473
This copies both the repository's revision data, and configuration information
2474
such as the make_working_trees setting.
2476
This is a destructive operation! Do not use it on existing
2479
:param revision_id: Only copy the content needed to construct
2480
revision_id and its parents.
2483
self.target.set_make_working_trees(self.source.make_working_trees())
2484
except NotImplementedError:
2486
# but don't bother fetching if we have the needed data now.
2487
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2488
self.target.has_revision(revision_id)):
2490
self.target.fetch(self.source, revision_id=revision_id)
2493
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2494
"""See InterRepository.fetch()."""
2495
from bzrlib.fetch import GenericRepoFetcher
2496
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2497
self.source, self.source._format, self.target,
2498
self.target._format)
2499
f = GenericRepoFetcher(to_repository=self.target,
2500
from_repository=self.source,
2501
last_revision=revision_id,
2502
pb=pb, find_ghosts=find_ghosts)
2503
return f.count_copied, f.failed_revisions
2506
class InterWeaveRepo(InterSameDataRepository):
2507
"""Optimised code paths between Weave based repositories.
2509
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2510
implemented lazy inter-object optimisation.
2514
def _get_repo_format_to_test(self):
2515
from bzrlib.repofmt import weaverepo
2516
return weaverepo.RepositoryFormat7()
2519
def is_compatible(source, target):
2520
"""Be compatible with known Weave formats.
2522
We don't test for the stores being of specific types because that
2523
could lead to confusing results, and there is no need to be
2526
from bzrlib.repofmt.weaverepo import (
2532
return (isinstance(source._format, (RepositoryFormat5,
2534
RepositoryFormat7)) and
2535
isinstance(target._format, (RepositoryFormat5,
2537
RepositoryFormat7)))
2538
except AttributeError:
2542
def copy_content(self, revision_id=None):
2543
"""See InterRepository.copy_content()."""
2544
# weave specific optimised path:
2546
self.target.set_make_working_trees(self.source.make_working_trees())
2547
except (errors.RepositoryUpgradeRequired, NotImplemented):
2549
# FIXME do not peek!
2550
if self.source._transport.listable():
2551
pb = ui.ui_factory.nested_progress_bar()
2553
self.target.weave_store.copy_all_ids(
2554
self.source.weave_store,
2556
from_transaction=self.source.get_transaction(),
2557
to_transaction=self.target.get_transaction())
2558
pb.update('copying inventory', 0, 1)
2559
self.target.control_weaves.copy_multi(
2560
self.source.control_weaves, ['inventory'],
2561
from_transaction=self.source.get_transaction(),
2562
to_transaction=self.target.get_transaction())
2563
self.target._revision_store.text_store.copy_all_ids(
2564
self.source._revision_store.text_store,
2569
self.target.fetch(self.source, revision_id=revision_id)
2572
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2573
"""See InterRepository.fetch()."""
2574
from bzrlib.fetch import GenericRepoFetcher
2575
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2576
self.source, self.source._format, self.target, self.target._format)
2577
f = GenericRepoFetcher(to_repository=self.target,
2578
from_repository=self.source,
2579
last_revision=revision_id,
2580
pb=pb, find_ghosts=find_ghosts)
2581
return f.count_copied, f.failed_revisions
2584
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2585
"""See InterRepository.missing_revision_ids()."""
2586
# we want all revisions to satisfy revision_id in source.
2587
# but we don't want to stat every file here and there.
2588
# we want then, all revisions other needs to satisfy revision_id
2589
# checked, but not those that we have locally.
2590
# so the first thing is to get a subset of the revisions to
2591
# satisfy revision_id in source, and then eliminate those that
2592
# we do already have.
2593
# this is slow on high latency connection to self, but as as this
2594
# disk format scales terribly for push anyway due to rewriting
2595
# inventory.weave, this is considered acceptable.
2597
if revision_id is not None:
2598
source_ids = self.source.get_ancestry(revision_id)
2599
if source_ids[0] is not None:
2600
raise AssertionError()
2603
source_ids = self.source._all_possible_ids()
2604
source_ids_set = set(source_ids)
2605
# source_ids is the worst possible case we may need to pull.
2606
# now we want to filter source_ids against what we actually
2607
# have in target, but don't try to check for existence where we know
2608
# we do not have a revision as that would be pointless.
2609
target_ids = set(self.target._all_possible_ids())
2610
possibly_present_revisions = target_ids.intersection(source_ids_set)
2611
actually_present_revisions = set(
2612
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2613
required_revisions = source_ids_set.difference(actually_present_revisions)
2614
if revision_id is not None:
2615
# we used get_ancestry to determine source_ids then we are assured all
2616
# revisions referenced are present as they are installed in topological order.
2617
# and the tip revision was validated by get_ancestry.
2618
result_set = required_revisions
2620
# if we just grabbed the possibly available ids, then
2621
# we only have an estimate of whats available and need to validate
2622
# that against the revision records.
2624
self.source._eliminate_revisions_not_present(required_revisions))
2625
return self.source.revision_ids_to_search_result(result_set)
2628
class InterKnitRepo(InterSameDataRepository):
2629
"""Optimised code paths between Knit based repositories."""
2632
def _get_repo_format_to_test(self):
2633
from bzrlib.repofmt import knitrepo
2634
return knitrepo.RepositoryFormatKnit1()
2637
def is_compatible(source, target):
2638
"""Be compatible with known Knit formats.
2640
We don't test for the stores being of specific types because that
2641
could lead to confusing results, and there is no need to be
2644
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2646
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2647
isinstance(target._format, RepositoryFormatKnit))
2648
except AttributeError:
2650
return are_knits and InterRepository._same_model(source, target)
2653
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2654
"""See InterRepository.fetch()."""
2655
from bzrlib.fetch import KnitRepoFetcher
2656
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2657
self.source, self.source._format, self.target, self.target._format)
2658
f = KnitRepoFetcher(to_repository=self.target,
2659
from_repository=self.source,
2660
last_revision=revision_id,
2661
pb=pb, find_ghosts=find_ghosts)
2662
return f.count_copied, f.failed_revisions
2665
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2666
"""See InterRepository.missing_revision_ids()."""
2667
if revision_id is not None:
2668
source_ids = self.source.get_ancestry(revision_id)
2669
if source_ids[0] is not None:
2670
raise AssertionError()
2673
source_ids = self.source.all_revision_ids()
2674
source_ids_set = set(source_ids)
2675
# source_ids is the worst possible case we may need to pull.
2676
# now we want to filter source_ids against what we actually
2677
# have in target, but don't try to check for existence where we know
2678
# we do not have a revision as that would be pointless.
2679
target_ids = set(self.target.all_revision_ids())
2680
possibly_present_revisions = target_ids.intersection(source_ids_set)
2681
actually_present_revisions = set(
2682
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2683
required_revisions = source_ids_set.difference(actually_present_revisions)
2684
if revision_id is not None:
2685
# we used get_ancestry to determine source_ids then we are assured all
2686
# revisions referenced are present as they are installed in topological order.
2687
# and the tip revision was validated by get_ancestry.
2688
result_set = required_revisions
2690
# if we just grabbed the possibly available ids, then
2691
# we only have an estimate of whats available and need to validate
2692
# that against the revision records.
2694
self.source._eliminate_revisions_not_present(required_revisions))
2695
return self.source.revision_ids_to_search_result(result_set)
2698
class InterPackRepo(InterSameDataRepository):
2699
"""Optimised code paths between Pack based repositories."""
2702
def _get_repo_format_to_test(self):
2703
from bzrlib.repofmt import pack_repo
2704
return pack_repo.RepositoryFormatKnitPack1()
2707
def is_compatible(source, target):
2708
"""Be compatible with known Pack formats.
2710
We don't test for the stores being of specific types because that
2711
could lead to confusing results, and there is no need to be
2714
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2716
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2717
isinstance(target._format, RepositoryFormatPack))
2718
except AttributeError:
2720
return are_packs and InterRepository._same_model(source, target)
2723
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2724
"""See InterRepository.fetch()."""
2725
from bzrlib.repofmt.pack_repo import Packer
2726
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2727
self.source, self.source._format, self.target, self.target._format)
2728
self.count_copied = 0
2729
if revision_id is None:
2731
# everything to do - use pack logic
2732
# to fetch from all packs to one without
2733
# inventory parsing etc, IFF nothing to be copied is in the target.
2735
source_revision_ids = frozenset(self.source.all_revision_ids())
2736
revision_ids = source_revision_ids - \
2737
frozenset(self.target.get_parent_map(source_revision_ids))
2738
revision_keys = [(revid,) for revid in revision_ids]
2739
index = self.target._pack_collection.revision_index.combined_index
2740
present_revision_ids = set(item[1][0] for item in
2741
index.iter_entries(revision_keys))
2742
revision_ids = set(revision_ids) - present_revision_ids
2743
# implementing the TODO will involve:
2744
# - detecting when all of a pack is selected
2745
# - avoiding as much as possible pre-selection, so the
2746
# more-core routines such as create_pack_from_packs can filter in
2747
# a just-in-time fashion. (though having a HEADS list on a
2748
# repository might make this a lot easier, because we could
2749
# sensibly detect 'new revisions' without doing a full index scan.
2750
elif _mod_revision.is_null(revision_id):
2755
revision_ids = self.search_missing_revision_ids(revision_id,
2756
find_ghosts=find_ghosts).get_keys()
2757
except errors.NoSuchRevision:
2758
raise errors.InstallFailed([revision_id])
2759
if len(revision_ids) == 0:
2761
packs = self.source._pack_collection.all_packs()
2762
pack = Packer(self.target._pack_collection, packs, '.fetch',
2763
revision_ids).pack()
2764
if pack is not None:
2765
self.target._pack_collection._save_pack_names()
2766
# Trigger an autopack. This may duplicate effort as we've just done
2767
# a pack creation, but for now it is simpler to think about as
2768
# 'upload data, then repack if needed'.
2769
self.target._pack_collection.autopack()
2770
return (pack.get_revision_count(), [])
2775
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2776
"""See InterRepository.missing_revision_ids().
2778
:param find_ghosts: Find ghosts throughout the ancestry of
2781
if not find_ghosts and revision_id is not None:
2782
return self._walk_to_common_revisions([revision_id])
2783
elif revision_id is not None:
2784
# Find ghosts: search for revisions pointing from one repository to
2785
# the other, and vice versa, anywhere in the history of revision_id.
2786
graph = self.target.get_graph(other_repository=self.source)
2787
searcher = graph._make_breadth_first_searcher([revision_id])
2791
next_revs, ghosts = searcher.next_with_ghosts()
2792
except StopIteration:
2794
if revision_id in ghosts:
2795
raise errors.NoSuchRevision(self.source, revision_id)
2796
found_ids.update(next_revs)
2797
found_ids.update(ghosts)
2798
found_ids = frozenset(found_ids)
2799
# Double query here: should be able to avoid this by changing the
2800
# graph api further.
2801
result_set = found_ids - frozenset(
2802
self.target.get_graph().get_parent_map(found_ids))
2804
source_ids = self.source.all_revision_ids()
2805
# source_ids is the worst possible case we may need to pull.
2806
# now we want to filter source_ids against what we actually
2807
# have in target, but don't try to check for existence where we know
2808
# we do not have a revision as that would be pointless.
2809
target_ids = set(self.target.all_revision_ids())
2810
result_set = set(source_ids).difference(target_ids)
2811
return self.source.revision_ids_to_search_result(result_set)
2814
class InterModel1and2(InterRepository):
2817
def _get_repo_format_to_test(self):
2821
def is_compatible(source, target):
2822
if not source.supports_rich_root() and target.supports_rich_root():
2828
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2829
"""See InterRepository.fetch()."""
2830
from bzrlib.fetch import Model1toKnit2Fetcher
2831
f = Model1toKnit2Fetcher(to_repository=self.target,
2832
from_repository=self.source,
2833
last_revision=revision_id,
2834
pb=pb, find_ghosts=find_ghosts)
2835
return f.count_copied, f.failed_revisions
2838
def copy_content(self, revision_id=None):
2839
"""Make a complete copy of the content in self into destination.
2841
This is a destructive operation! Do not use it on existing
2844
:param revision_id: Only copy the content needed to construct
2845
revision_id and its parents.
2848
self.target.set_make_working_trees(self.source.make_working_trees())
2849
except NotImplementedError:
2851
# but don't bother fetching if we have the needed data now.
2852
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2853
self.target.has_revision(revision_id)):
2855
self.target.fetch(self.source, revision_id=revision_id)
2858
class InterKnit1and2(InterKnitRepo):
2861
def _get_repo_format_to_test(self):
2865
def is_compatible(source, target):
2866
"""Be compatible with Knit1 source and Knit3 target"""
2867
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2869
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2870
RepositoryFormatKnit3)
2871
from bzrlib.repofmt.pack_repo import (
2872
RepositoryFormatKnitPack1,
2873
RepositoryFormatKnitPack3,
2874
RepositoryFormatPackDevelopment0,
2875
RepositoryFormatPackDevelopment0Subtree,
2878
RepositoryFormatKnit1,
2879
RepositoryFormatKnitPack1,
2880
RepositoryFormatPackDevelopment0,
2883
RepositoryFormatKnit3,
2884
RepositoryFormatKnitPack3,
2885
RepositoryFormatPackDevelopment0Subtree,
2887
return (isinstance(source._format, nosubtrees) and
2888
isinstance(target._format, subtrees))
2889
except AttributeError:
2893
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2894
"""See InterRepository.fetch()."""
2895
from bzrlib.fetch import Knit1to2Fetcher
2896
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2897
self.source, self.source._format, self.target,
2898
self.target._format)
2899
f = Knit1to2Fetcher(to_repository=self.target,
2900
from_repository=self.source,
2901
last_revision=revision_id,
2902
pb=pb, find_ghosts=find_ghosts)
2903
return f.count_copied, f.failed_revisions
2906
class InterDifferingSerializer(InterKnitRepo):
2909
def _get_repo_format_to_test(self):
2913
def is_compatible(source, target):
2914
"""Be compatible with Knit2 source and Knit3 target"""
2915
if source.supports_rich_root() != target.supports_rich_root():
2917
# Ideally, we'd support fetching if the source had no tree references
2918
# even if it supported them...
2919
if (getattr(source, '_format.supports_tree_reference', False) and
2920
not getattr(target, '_format.supports_tree_reference', False)):
2925
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2926
"""See InterRepository.fetch()."""
2927
revision_ids = self.target.search_missing_revision_ids(self.source,
2928
revision_id, find_ghosts=find_ghosts).get_keys()
2929
revision_ids = tsort.topo_sort(
2930
self.source.get_graph().get_parent_map(revision_ids))
2931
def revisions_iterator():
2932
for current_revision_id in revision_ids:
2933
revision = self.source.get_revision(current_revision_id)
2934
tree = self.source.revision_tree(current_revision_id)
2936
signature = self.source.get_signature_text(
2937
current_revision_id)
2938
except errors.NoSuchRevision:
2940
yield revision, tree, signature
2942
my_pb = ui.ui_factory.nested_progress_bar()
2947
install_revisions(self.target, revisions_iterator(),
2948
len(revision_ids), pb)
2950
if my_pb is not None:
2952
return len(revision_ids), 0
2955
class InterRemoteToOther(InterRepository):
2957
def __init__(self, source, target):
2958
InterRepository.__init__(self, source, target)
2959
self._real_inter = None
2962
def is_compatible(source, target):
2963
if not isinstance(source, remote.RemoteRepository):
2965
# Is source's model compatible with target's model?
2966
source._ensure_real()
2967
real_source = source._real_repository
2968
if isinstance(real_source, remote.RemoteRepository):
2969
raise NotImplementedError(
2970
"We don't support remote repos backed by remote repos yet.")
2971
return InterRepository._same_model(real_source, target)
2974
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2975
"""See InterRepository.fetch()."""
2976
from bzrlib.fetch import RemoteToOtherFetcher
2977
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2978
self.source, self.target, self.target._format)
2979
# TODO: jam 20070210 This should be an assert, not a translate
2980
revision_id = osutils.safe_revision_id(revision_id)
2981
f = RemoteToOtherFetcher(to_repository=self.target,
2982
from_repository=self.source,
2983
last_revision=revision_id,
2984
pb=pb, find_ghosts=find_ghosts)
2985
return f.count_copied, f.failed_revisions
2988
def _get_repo_format_to_test(self):
2992
class InterOtherToRemote(InterRepository):
2994
def __init__(self, source, target):
2995
InterRepository.__init__(self, source, target)
2996
self._real_inter = None
2999
def is_compatible(source, target):
3000
if isinstance(target, remote.RemoteRepository):
3004
def _ensure_real_inter(self):
3005
if self._real_inter is None:
3006
self.target._ensure_real()
3007
real_target = self.target._real_repository
3008
self._real_inter = InterRepository.get(self.source, real_target)
3010
def copy_content(self, revision_id=None):
3011
self._ensure_real_inter()
3012
self._real_inter.copy_content(revision_id=revision_id)
3014
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
3015
self._ensure_real_inter()
3016
self._real_inter.fetch(revision_id=revision_id, pb=pb,
3017
find_ghosts=find_ghosts)
3020
def _get_repo_format_to_test(self):
3024
InterRepository.register_optimiser(InterDifferingSerializer)
3025
InterRepository.register_optimiser(InterSameDataRepository)
3026
InterRepository.register_optimiser(InterWeaveRepo)
3027
InterRepository.register_optimiser(InterKnitRepo)
3028
InterRepository.register_optimiser(InterModel1and2)
3029
InterRepository.register_optimiser(InterKnit1and2)
3030
InterRepository.register_optimiser(InterPackRepo)
3031
InterRepository.register_optimiser(InterRemoteToOther)
3032
InterRepository.register_optimiser(InterOtherToRemote)
3035
class CopyConverter(object):
3036
"""A repository conversion tool which just performs a copy of the content.
3038
This is slow but quite reliable.
3041
def __init__(self, target_format):
3042
"""Create a CopyConverter.
3044
:param target_format: The format the resulting repository should be.
3046
self.target_format = target_format
3048
def convert(self, repo, pb):
3049
"""Perform the conversion of to_convert, giving feedback via pb.
3051
:param to_convert: The disk object to convert.
3052
:param pb: a progress bar to use for progress information.
3057
# this is only useful with metadir layouts - separated repo content.
3058
# trigger an assertion if not such
3059
repo._format.get_format_string()
3060
self.repo_dir = repo.bzrdir
3061
self.step('Moving repository to repository.backup')
3062
self.repo_dir.transport.move('repository', 'repository.backup')
3063
backup_transport = self.repo_dir.transport.clone('repository.backup')
3064
repo._format.check_conversion_target(self.target_format)
3065
self.source_repo = repo._format.open(self.repo_dir,
3067
_override_transport=backup_transport)
3068
self.step('Creating new repository')
3069
converted = self.target_format.initialize(self.repo_dir,
3070
self.source_repo.is_shared())
3071
converted.lock_write()
3073
self.step('Copying content into repository.')
3074
self.source_repo.copy_content_into(converted)
3077
self.step('Deleting old repository content.')
3078
self.repo_dir.transport.delete_tree('repository.backup')
3079
self.pb.note('repository converted')
3081
def step(self, message):
3082
"""Update the pb by a step."""
3084
self.pb.update(message, self.count, self.total)
3096
def _unescaper(match, _map=_unescape_map):
3097
code = match.group(1)
3101
if not code.startswith('#'):
3103
return unichr(int(code[1:])).encode('utf8')
3109
def _unescape_xml(data):
3110
"""Unescape predefined XML entities in a string of data."""
3112
if _unescape_re is None:
3113
_unescape_re = re.compile('\&([^;]*);')
3114
return _unescape_re.sub(_unescaper, data)
3117
class _VersionedFileChecker(object):
3119
def __init__(self, repository):
3120
self.repository = repository
3121
self.text_index = self.repository._generate_text_key_index()
3123
def calculate_file_version_parents(self, revision_id, file_id):
3124
"""Calculate the correct parents for a file version according to
3127
parent_keys = self.text_index[(file_id, revision_id)]
3128
if parent_keys == [_mod_revision.NULL_REVISION]:
3130
# strip the file_id, for the weave api
3131
return tuple([revision_id for file_id, revision_id in parent_keys])
3133
def check_file_version_parents(self, weave, file_id):
3134
"""Check the parents stored in a versioned file are correct.
3136
It also detects file versions that are not referenced by their
3137
corresponding revision's inventory.
3139
:returns: A tuple of (wrong_parents, dangling_file_versions).
3140
wrong_parents is a dict mapping {revision_id: (stored_parents,
3141
correct_parents)} for each revision_id where the stored parents
3142
are not correct. dangling_file_versions is a set of (file_id,
3143
revision_id) tuples for versions that are present in this versioned
3144
file, but not used by the corresponding inventory.
3147
unused_versions = set()
3148
versions = weave.versions()
3149
parent_map = weave.get_parent_map(versions)
3150
for num, revision_id in enumerate(versions):
3152
correct_parents = self.calculate_file_version_parents(
3153
revision_id, file_id)
3155
# The version is not part of the used keys.
3156
unused_versions.add(revision_id)
3159
knit_parents = tuple(parent_map[revision_id])
3160
except errors.RevisionNotPresent:
3162
if correct_parents != knit_parents:
3163
wrong_parents[revision_id] = (knit_parents, correct_parents)
3164
return wrong_parents, unused_versions
3167
def _old_get_graph(repository, revision_id):
3168
"""DO NOT USE. That is all. I'm serious."""
3169
graph = repository.get_graph()
3170
revision_graph = dict(((key, value) for key, value in
3171
graph.iter_ancestry([revision_id]) if value is not None))
3172
return _strip_NULL_ghosts(revision_graph)
3175
def _strip_NULL_ghosts(revision_graph):
3176
"""Also don't use this. more compatibility code for unmigrated clients."""
3177
# Filter ghosts, and null:
3178
if _mod_revision.NULL_REVISION in revision_graph:
3179
del revision_graph[_mod_revision.NULL_REVISION]
3180
for key, parents in revision_graph.items():
3181
revision_graph[key] = tuple(parent for parent in parents if parent
3183
return revision_graph
added
removed
bzrlib/repository.py
