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(), """
40
revision as _mod_revision,
46
from bzrlib.bundle import serializer
47
from bzrlib.revisiontree import RevisionTree
48
from bzrlib.store.versioned import VersionedFileStore
49
from bzrlib.store.text import TextStore
50
from bzrlib.testament import Testament
51
from bzrlib.util import bencode
54
from bzrlib.decorators import needs_read_lock, needs_write_lock
55
from bzrlib.inter import InterObject
56
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
57
from bzrlib.symbol_versioning import (
60
from bzrlib.trace import mutter, mutter_callsite, note, warning
63
# Old formats display a warning, but only once
64
_deprecation_warning_done = False
67
class CommitBuilder(object):
68
"""Provides an interface to build up a commit.
70
This allows describing a tree to be committed without needing to
71
know the internals of the format of the repository.
74
# all clients should supply tree roots.
75
record_root_entry = True
76
# the default CommitBuilder does not manage trees whose root is versioned.
77
_versioned_root = False
79
def __init__(self, repository, parents, config, timestamp=None,
80
timezone=None, committer=None, revprops=None,
82
"""Initiate a CommitBuilder.
84
:param repository: Repository to commit to.
85
:param parents: Revision ids of the parents of the new revision.
86
:param config: Configuration to use.
87
:param timestamp: Optional timestamp recorded for commit.
88
:param timezone: Optional timezone for timestamp.
89
:param committer: Optional committer to set for commit.
90
:param revprops: Optional dictionary of revision properties.
91
:param revision_id: Optional revision id.
96
self._committer = self._config.username()
98
self._committer = committer
100
self.new_inventory = Inventory(None)
101
self._new_revision_id = revision_id
102
self.parents = parents
103
self.repository = repository
106
if revprops is not None:
107
self._revprops.update(revprops)
109
if timestamp is None:
110
timestamp = time.time()
111
# Restrict resolution to 1ms
112
self._timestamp = round(timestamp, 3)
115
self._timezone = osutils.local_time_offset()
117
self._timezone = int(timezone)
119
self._generate_revision_if_needed()
120
self.__heads = graph.HeadsCache(repository.get_graph()).heads
122
def commit(self, message):
123
"""Make the actual commit.
125
:return: The revision id of the recorded revision.
127
rev = _mod_revision.Revision(
128
timestamp=self._timestamp,
129
timezone=self._timezone,
130
committer=self._committer,
132
inventory_sha1=self.inv_sha1,
133
revision_id=self._new_revision_id,
134
properties=self._revprops)
135
rev.parent_ids = self.parents
136
self.repository.add_revision(self._new_revision_id, rev,
137
self.new_inventory, self._config)
138
self.repository.commit_write_group()
139
return self._new_revision_id
142
"""Abort the commit that is being built.
144
self.repository.abort_write_group()
146
def revision_tree(self):
147
"""Return the tree that was just committed.
149
After calling commit() this can be called to get a RevisionTree
150
representing the newly committed tree. This is preferred to
151
calling Repository.revision_tree() because that may require
152
deserializing the inventory, while we already have a copy in
155
return RevisionTree(self.repository, self.new_inventory,
156
self._new_revision_id)
158
def finish_inventory(self):
159
"""Tell the builder that the inventory is finished."""
160
if self.new_inventory.root is None:
161
raise AssertionError('Root entry should be supplied to'
162
' record_entry_contents, as of bzr 0.10.',
163
DeprecationWarning, stacklevel=2)
164
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
165
self.new_inventory.revision_id = self._new_revision_id
166
self.inv_sha1 = self.repository.add_inventory(
167
self._new_revision_id,
172
def _gen_revision_id(self):
173
"""Return new revision-id."""
174
return generate_ids.gen_revision_id(self._config.username(),
177
def _generate_revision_if_needed(self):
178
"""Create a revision id if None was supplied.
180
If the repository can not support user-specified revision ids
181
they should override this function and raise CannotSetRevisionId
182
if _new_revision_id is not None.
184
:raises: CannotSetRevisionId
186
if self._new_revision_id is None:
187
self._new_revision_id = self._gen_revision_id()
188
self.random_revid = True
190
self.random_revid = False
192
def _heads(self, file_id, revision_ids):
193
"""Calculate the graph heads for revision_ids in the graph of file_id.
195
This can use either a per-file graph or a global revision graph as we
196
have an identity relationship between the two graphs.
198
return self.__heads(revision_ids)
200
def _check_root(self, ie, parent_invs, tree):
201
"""Helper for record_entry_contents.
203
:param ie: An entry being added.
204
:param parent_invs: The inventories of the parent revisions of the
206
:param tree: The tree that is being committed.
208
# In this revision format, root entries have no knit or weave When
209
# serializing out to disk and back in root.revision is always
211
ie.revision = self._new_revision_id
213
def _get_delta(self, ie, basis_inv, path):
214
"""Get a delta against the basis inventory for ie."""
215
if ie.file_id not in basis_inv:
217
return (None, path, ie.file_id, ie)
218
elif ie != basis_inv[ie.file_id]:
220
# TODO: avoid tis id2path call.
221
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
226
def record_entry_contents(self, ie, parent_invs, path, tree,
228
"""Record the content of ie from tree into the commit if needed.
230
Side effect: sets ie.revision when unchanged
232
:param ie: An inventory entry present in the commit.
233
:param parent_invs: The inventories of the parent revisions of the
235
:param path: The path the entry is at in the tree.
236
:param tree: The tree which contains this entry and should be used to
238
:param content_summary: Summary data from the tree about the paths
239
content - stat, length, exec, sha/link target. This is only
240
accessed when the entry has a revision of None - that is when it is
241
a candidate to commit.
242
:return: A tuple (change_delta, version_recorded). change_delta is
243
an inventory_delta change for this entry against the basis tree of
244
the commit, or None if no change occured against the basis tree.
245
version_recorded is True if a new version of the entry has been
246
recorded. For instance, committing a merge where a file was only
247
changed on the other side will return (delta, False).
249
if self.new_inventory.root is None:
250
if ie.parent_id is not None:
251
raise errors.RootMissing()
252
self._check_root(ie, parent_invs, tree)
253
if ie.revision is None:
254
kind = content_summary[0]
256
# ie is carried over from a prior commit
258
# XXX: repository specific check for nested tree support goes here - if
259
# the repo doesn't want nested trees we skip it ?
260
if (kind == 'tree-reference' and
261
not self.repository._format.supports_tree_reference):
262
# mismatch between commit builder logic and repository:
263
# this needs the entry creation pushed down into the builder.
264
raise NotImplementedError('Missing repository subtree support.')
265
self.new_inventory.add(ie)
267
# TODO: slow, take it out of the inner loop.
269
basis_inv = parent_invs[0]
271
basis_inv = Inventory(root_id=None)
273
# ie.revision is always None if the InventoryEntry is considered
274
# for committing. We may record the previous parents revision if the
275
# content is actually unchanged against a sole head.
276
if ie.revision is not None:
277
if not self._versioned_root and path == '':
278
# repositories that do not version the root set the root's
279
# revision to the new commit even when no change occurs, and
280
# this masks when a change may have occurred against the basis,
281
# so calculate if one happened.
282
if ie.file_id in basis_inv:
283
delta = (basis_inv.id2path(ie.file_id), path,
287
delta = (None, path, ie.file_id, ie)
290
# we don't need to commit this, because the caller already
291
# determined that an existing revision of this file is
293
return None, (ie.revision == self._new_revision_id)
294
# XXX: Friction: parent_candidates should return a list not a dict
295
# so that we don't have to walk the inventories again.
296
parent_candiate_entries = ie.parent_candidates(parent_invs)
297
head_set = self._heads(ie.file_id, parent_candiate_entries.keys())
299
for inv in parent_invs:
300
if ie.file_id in inv:
301
old_rev = inv[ie.file_id].revision
302
if old_rev in head_set:
303
heads.append(inv[ie.file_id].revision)
304
head_set.remove(inv[ie.file_id].revision)
307
# now we check to see if we need to write a new record to the
309
# We write a new entry unless there is one head to the ancestors, and
310
# the kind-derived content is unchanged.
312
# Cheapest check first: no ancestors, or more the one head in the
313
# ancestors, we write a new node.
317
# There is a single head, look it up for comparison
318
parent_entry = parent_candiate_entries[heads[0]]
319
# if the non-content specific data has changed, we'll be writing a
321
if (parent_entry.parent_id != ie.parent_id or
322
parent_entry.name != ie.name):
324
# now we need to do content specific checks:
326
# if the kind changed the content obviously has
327
if kind != parent_entry.kind:
330
if content_summary[2] is None:
331
raise ValueError("Files must not have executable = None")
333
if (# if the file length changed we have to store:
334
parent_entry.text_size != content_summary[1] or
335
# if the exec bit has changed we have to store:
336
parent_entry.executable != content_summary[2]):
338
elif parent_entry.text_sha1 == content_summary[3]:
339
# all meta and content is unchanged (using a hash cache
340
# hit to check the sha)
341
ie.revision = parent_entry.revision
342
ie.text_size = parent_entry.text_size
343
ie.text_sha1 = parent_entry.text_sha1
344
ie.executable = parent_entry.executable
345
return self._get_delta(ie, basis_inv, path), False
347
# Either there is only a hash change(no hash cache entry,
348
# or same size content change), or there is no change on
350
# Provide the parent's hash to the store layer, so that the
351
# content is unchanged we will not store a new node.
352
nostore_sha = parent_entry.text_sha1
354
# We want to record a new node regardless of the presence or
355
# absence of a content change in the file.
357
ie.executable = content_summary[2]
358
lines = tree.get_file(ie.file_id, path).readlines()
360
ie.text_sha1, ie.text_size = self._add_text_to_weave(
361
ie.file_id, lines, heads, nostore_sha)
362
except errors.ExistingContent:
363
# Turns out that the file content was unchanged, and we were
364
# only going to store a new node if it was changed. Carry over
366
ie.revision = parent_entry.revision
367
ie.text_size = parent_entry.text_size
368
ie.text_sha1 = parent_entry.text_sha1
369
ie.executable = parent_entry.executable
370
return self._get_delta(ie, basis_inv, path), False
371
elif kind == 'directory':
373
# all data is meta here, nothing specific to directory, so
375
ie.revision = parent_entry.revision
376
return self._get_delta(ie, basis_inv, path), False
378
self._add_text_to_weave(ie.file_id, lines, heads, None)
379
elif kind == 'symlink':
380
current_link_target = content_summary[3]
382
# symlink target is not generic metadata, check if it has
384
if current_link_target != parent_entry.symlink_target:
387
# unchanged, carry over.
388
ie.revision = parent_entry.revision
389
ie.symlink_target = parent_entry.symlink_target
390
return self._get_delta(ie, basis_inv, path), False
391
ie.symlink_target = current_link_target
393
self._add_text_to_weave(ie.file_id, lines, heads, None)
394
elif kind == 'tree-reference':
396
if content_summary[3] != parent_entry.reference_revision:
399
# unchanged, carry over.
400
ie.reference_revision = parent_entry.reference_revision
401
ie.revision = parent_entry.revision
402
return self._get_delta(ie, basis_inv, path), False
403
ie.reference_revision = content_summary[3]
405
self._add_text_to_weave(ie.file_id, lines, heads, None)
407
raise NotImplementedError('unknown kind')
408
ie.revision = self._new_revision_id
409
return self._get_delta(ie, basis_inv, path), True
411
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
412
versionedfile = self.repository.weave_store.get_weave_or_empty(
413
file_id, self.repository.get_transaction())
414
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
415
# than add_lines, and allows committing when a parent is ghosted for
417
# Note: as we read the content directly from the tree, we know its not
418
# been turned into unicode or badly split - but a broken tree
419
# implementation could give us bad output from readlines() so this is
420
# not a guarantee of safety. What would be better is always checking
421
# the content during test suite execution. RBC 20070912
422
return versionedfile.add_lines_with_ghosts(
423
self._new_revision_id, parents, new_lines,
424
nostore_sha=nostore_sha, random_id=self.random_revid,
425
check_content=False)[0:2]
428
class RootCommitBuilder(CommitBuilder):
429
"""This commitbuilder actually records the root id"""
431
# the root entry gets versioned properly by this builder.
432
_versioned_root = True
434
def _check_root(self, ie, parent_invs, tree):
435
"""Helper for record_entry_contents.
437
:param ie: An entry being added.
438
:param parent_invs: The inventories of the parent revisions of the
440
:param tree: The tree that is being committed.
444
######################################################################
447
class Repository(object):
448
"""Repository holding history for one or more branches.
450
The repository holds and retrieves historical information including
451
revisions and file history. It's normally accessed only by the Branch,
452
which views a particular line of development through that history.
454
The Repository builds on top of Stores and a Transport, which respectively
455
describe the disk data format and the way of accessing the (possibly
459
# What class to use for a CommitBuilder. Often its simpler to change this
460
# in a Repository class subclass rather than to override
461
# get_commit_builder.
462
_commit_builder_class = CommitBuilder
463
# The search regex used by xml based repositories to determine what things
464
# where changed in a single commit.
465
_file_ids_altered_regex = lazy_regex.lazy_compile(
466
r'file_id="(?P<file_id>[^"]+)"'
467
r'.* revision="(?P<revision_id>[^"]+)"'
470
def abort_write_group(self):
471
"""Commit the contents accrued within the current write group.
473
:seealso: start_write_group.
475
if self._write_group is not self.get_transaction():
476
# has an unlock or relock occured ?
477
raise errors.BzrError('mismatched lock context and write group.')
478
self._abort_write_group()
479
self._write_group = None
481
def _abort_write_group(self):
482
"""Template method for per-repository write group cleanup.
484
This is called during abort before the write group is considered to be
485
finished and should cleanup any internal state accrued during the write
486
group. There is no requirement that data handed to the repository be
487
*not* made available - this is not a rollback - but neither should any
488
attempt be made to ensure that data added is fully commited. Abort is
489
invoked when an error has occured so futher disk or network operations
490
may not be possible or may error and if possible should not be
494
def add_inventory(self, revision_id, inv, parents):
495
"""Add the inventory inv to the repository as revision_id.
497
:param parents: The revision ids of the parents that revision_id
498
is known to have and are in the repository already.
500
:returns: The validator(which is a sha1 digest, though what is sha'd is
501
repository format specific) of the serialized inventory.
503
if not self.is_in_write_group():
504
raise AssertionError("%r not in write group" % (self,))
505
_mod_revision.check_not_reserved_id(revision_id)
506
if not (inv.revision_id is None or inv.revision_id == revision_id):
507
raise AssertionError(
508
"Mismatch between inventory revision"
509
" id and insertion revid (%r, %r)"
510
% (inv.revision_id, revision_id))
512
raise AssertionError()
513
inv_lines = self._serialise_inventory_to_lines(inv)
514
inv_vf = self.get_inventory_weave()
515
return self._inventory_add_lines(inv_vf, revision_id, parents,
516
inv_lines, check_content=False)
518
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
520
"""Store lines in inv_vf and return the sha1 of the inventory."""
522
for parent in parents:
524
final_parents.append(parent)
525
return inv_vf.add_lines(revision_id, final_parents, lines,
526
check_content=check_content)[0]
528
def add_revision(self, revision_id, rev, inv=None, config=None):
529
"""Add rev to the revision store as revision_id.
531
:param revision_id: the revision id to use.
532
:param rev: The revision object.
533
:param inv: The inventory for the revision. if None, it will be looked
534
up in the inventory storer
535
:param config: If None no digital signature will be created.
536
If supplied its signature_needed method will be used
537
to determine if a signature should be made.
539
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
541
_mod_revision.check_not_reserved_id(revision_id)
542
if config is not None and config.signature_needed():
544
inv = self.get_inventory(revision_id)
545
plaintext = Testament(rev, inv).as_short_text()
546
self.store_revision_signature(
547
gpg.GPGStrategy(config), plaintext, revision_id)
548
if not revision_id in self.get_inventory_weave():
550
raise errors.WeaveRevisionNotPresent(revision_id,
551
self.get_inventory_weave())
553
# yes, this is not suitable for adding with ghosts.
554
rev.inventory_sha1 = self.add_inventory(revision_id, inv,
556
self._revision_store.add_revision(rev, self.get_transaction())
558
def _add_revision_text(self, revision_id, text):
559
revision = self._revision_store._serializer.read_revision_from_string(
561
self._revision_store._add_revision(revision, StringIO(text),
562
self.get_transaction())
564
def all_revision_ids(self):
565
"""Returns a list of all the revision ids in the repository.
567
This is deprecated because code should generally work on the graph
568
reachable from a particular revision, and ignore any other revisions
569
that might be present. There is no direct replacement method.
571
if 'evil' in debug.debug_flags:
572
mutter_callsite(2, "all_revision_ids is linear with history.")
573
return self._all_revision_ids()
575
def _all_revision_ids(self):
576
"""Returns a list of all the revision ids in the repository.
578
These are in as much topological order as the underlying store can
581
raise NotImplementedError(self._all_revision_ids)
583
def break_lock(self):
584
"""Break a lock if one is present from another instance.
586
Uses the ui factory to ask for confirmation if the lock may be from
589
self.control_files.break_lock()
592
def _eliminate_revisions_not_present(self, revision_ids):
593
"""Check every revision id in revision_ids to see if we have it.
595
Returns a set of the present revisions.
598
graph = self.get_graph()
599
parent_map = graph.get_parent_map(revision_ids)
600
# The old API returned a list, should this actually be a set?
601
return parent_map.keys()
604
def create(a_bzrdir):
605
"""Construct the current default format repository in a_bzrdir."""
606
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
608
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
609
"""instantiate a Repository.
611
:param _format: The format of the repository on disk.
612
:param a_bzrdir: The BzrDir of the repository.
614
In the future we will have a single api for all stores for
615
getting file texts, inventories and revisions, then
616
this construct will accept instances of those things.
618
super(Repository, self).__init__()
619
self._format = _format
620
# the following are part of the public API for Repository:
621
self.bzrdir = a_bzrdir
622
self.control_files = control_files
623
self._revision_store = _revision_store
624
# backwards compatibility
625
self.weave_store = text_store
627
self._reconcile_does_inventory_gc = True
628
self._reconcile_fixes_text_parents = False
629
self._reconcile_backsup_inventory = True
630
# not right yet - should be more semantically clear ?
632
self.control_store = control_store
633
self.control_weaves = control_store
634
# TODO: make sure to construct the right store classes, etc, depending
635
# on whether escaping is required.
636
self._warn_if_deprecated()
637
self._write_group = None
638
self.base = control_files._transport.base
641
return '%s(%r)' % (self.__class__.__name__,
644
def has_same_location(self, other):
645
"""Returns a boolean indicating if this repository is at the same
646
location as another repository.
648
This might return False even when two repository objects are accessing
649
the same physical repository via different URLs.
651
if self.__class__ is not other.__class__:
653
return (self.control_files._transport.base ==
654
other.control_files._transport.base)
656
def is_in_write_group(self):
657
"""Return True if there is an open write group.
659
:seealso: start_write_group.
661
return self._write_group is not None
664
return self.control_files.is_locked()
666
def is_write_locked(self):
667
"""Return True if this object is write locked."""
668
return self.is_locked() and self.control_files._lock_mode == 'w'
670
def lock_write(self, token=None):
671
"""Lock this repository for writing.
673
This causes caching within the repository obejct to start accumlating
674
data during reads, and allows a 'write_group' to be obtained. Write
675
groups must be used for actual data insertion.
677
:param token: if this is already locked, then lock_write will fail
678
unless the token matches the existing lock.
679
:returns: a token if this instance supports tokens, otherwise None.
680
:raises TokenLockingNotSupported: when a token is given but this
681
instance doesn't support using token locks.
682
:raises MismatchedToken: if the specified token doesn't match the token
683
of the existing lock.
684
:seealso: start_write_group.
686
A token should be passed in if you know that you have locked the object
687
some other way, and need to synchronise this object's state with that
690
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
692
result = self.control_files.lock_write(token=token)
697
self.control_files.lock_read()
700
def get_physical_lock_status(self):
701
return self.control_files.get_physical_lock_status()
703
def leave_lock_in_place(self):
704
"""Tell this repository not to release the physical lock when this
707
If lock_write doesn't return a token, then this method is not supported.
709
self.control_files.leave_in_place()
711
def dont_leave_lock_in_place(self):
712
"""Tell this repository to release the physical lock when this
713
object is unlocked, even if it didn't originally acquire it.
715
If lock_write doesn't return a token, then this method is not supported.
717
self.control_files.dont_leave_in_place()
720
def gather_stats(self, revid=None, committers=None):
721
"""Gather statistics from a revision id.
723
:param revid: The revision id to gather statistics from, if None, then
724
no revision specific statistics are gathered.
725
:param committers: Optional parameter controlling whether to grab
726
a count of committers from the revision specific statistics.
727
:return: A dictionary of statistics. Currently this contains:
728
committers: The number of committers if requested.
729
firstrev: A tuple with timestamp, timezone for the penultimate left
730
most ancestor of revid, if revid is not the NULL_REVISION.
731
latestrev: A tuple with timestamp, timezone for revid, if revid is
732
not the NULL_REVISION.
733
revisions: The total revision count in the repository.
734
size: An estimate disk size of the repository in bytes.
737
if revid and committers:
738
result['committers'] = 0
739
if revid and revid != _mod_revision.NULL_REVISION:
741
all_committers = set()
742
revisions = self.get_ancestry(revid)
743
# pop the leading None
745
first_revision = None
747
# ignore the revisions in the middle - just grab first and last
748
revisions = revisions[0], revisions[-1]
749
for revision in self.get_revisions(revisions):
750
if not first_revision:
751
first_revision = revision
753
all_committers.add(revision.committer)
754
last_revision = revision
756
result['committers'] = len(all_committers)
757
result['firstrev'] = (first_revision.timestamp,
758
first_revision.timezone)
759
result['latestrev'] = (last_revision.timestamp,
760
last_revision.timezone)
762
# now gather global repository information
763
if self.bzrdir.root_transport.listable():
764
c, t = self._revision_store.total_size(self.get_transaction())
765
result['revisions'] = c
769
def find_branches(self, using=False):
770
"""Find branches underneath this repository.
772
This will include branches inside other branches.
774
:param using: If True, list only branches using this repository.
776
if using and not self.is_shared():
778
return [self.bzrdir.open_branch()]
779
except errors.NotBranchError:
781
class Evaluator(object):
784
self.first_call = True
786
def __call__(self, bzrdir):
787
# On the first call, the parameter is always the bzrdir
788
# containing the current repo.
789
if not self.first_call:
791
repository = bzrdir.open_repository()
792
except errors.NoRepositoryPresent:
795
return False, (None, repository)
796
self.first_call = False
798
value = (bzrdir.open_branch(), None)
799
except errors.NotBranchError:
804
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
805
self.bzrdir.root_transport, evaluate=Evaluator()):
806
if branch is not None:
807
branches.append(branch)
808
if not using and repository is not None:
809
branches.extend(repository.find_branches())
812
def get_data_stream(self, revision_ids):
813
raise NotImplementedError(self.get_data_stream)
815
def get_data_stream_for_search(self, search_result):
816
"""Get a data stream that can be inserted to a repository.
818
:param search_result: A bzrlib.graph.SearchResult selecting the
820
:return: A data stream that can be inserted into a repository using
823
raise NotImplementedError(self.get_data_stream_for_search)
825
def insert_data_stream(self, stream):
826
"""XXX What does this really do?
828
Is it a substitute for fetch?
829
Should it manage its own write group ?
831
for item_key, bytes in stream:
832
if item_key[0] == 'file':
833
(file_id,) = item_key[1:]
834
knit = self.weave_store.get_weave_or_empty(
835
file_id, self.get_transaction())
836
elif item_key == ('inventory',):
837
knit = self.get_inventory_weave()
838
elif item_key == ('revisions',):
839
knit = self._revision_store.get_revision_file(
840
self.get_transaction())
841
elif item_key == ('signatures',):
842
knit = self._revision_store.get_signature_file(
843
self.get_transaction())
845
raise errors.RepositoryDataStreamError(
846
"Unrecognised data stream key '%s'" % (item_key,))
847
decoded_list = bencode.bdecode(bytes)
848
format = decoded_list.pop(0)
851
for version, options, parents, some_bytes in decoded_list:
852
data_list.append((version, options, len(some_bytes), parents))
853
knit_bytes += some_bytes
854
buffer = StringIO(knit_bytes)
855
def reader_func(count):
859
return buffer.read(count)
860
knit.insert_data_stream(
861
(format, data_list, reader_func))
864
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
865
"""Return the revision ids that other has that this does not.
867
These are returned in topological order.
869
revision_id: only return revision ids included by revision_id.
871
return InterRepository.get(other, self).search_missing_revision_ids(
872
revision_id, find_ghosts)
874
@deprecated_method(symbol_versioning.one_two)
876
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
877
"""Return the revision ids that other has that this does not.
879
These are returned in topological order.
881
revision_id: only return revision ids included by revision_id.
883
keys = self.search_missing_revision_ids(
884
other, revision_id, find_ghosts).get_keys()
887
parents = other.get_graph().get_parent_map(keys)
890
return tsort.topo_sort(parents)
894
"""Open the repository rooted at base.
896
For instance, if the repository is at URL/.bzr/repository,
897
Repository.open(URL) -> a Repository instance.
899
control = bzrdir.BzrDir.open(base)
900
return control.open_repository()
902
def copy_content_into(self, destination, revision_id=None):
903
"""Make a complete copy of the content in self into destination.
905
This is a destructive operation! Do not use it on existing
908
return InterRepository.get(self, destination).copy_content(revision_id)
910
def commit_write_group(self):
911
"""Commit the contents accrued within the current write group.
913
:seealso: start_write_group.
915
if self._write_group is not self.get_transaction():
916
# has an unlock or relock occured ?
917
raise errors.BzrError('mismatched lock context %r and '
919
(self.get_transaction(), self._write_group))
920
self._commit_write_group()
921
self._write_group = None
923
def _commit_write_group(self):
924
"""Template method for per-repository write group cleanup.
926
This is called before the write group is considered to be
927
finished and should ensure that all data handed to the repository
928
for writing during the write group is safely committed (to the
929
extent possible considering file system caching etc).
932
def fetch(self, source, revision_id=None, pb=None, find_ghosts=False):
933
"""Fetch the content required to construct revision_id from source.
935
If revision_id is None all content is copied.
936
:param find_ghosts: Find and copy revisions in the source that are
937
ghosts in the target (and not reachable directly by walking out to
938
the first-present revision in target from revision_id).
940
# fast path same-url fetch operations
941
if self.has_same_location(source):
942
# check that last_revision is in 'from' and then return a
944
if (revision_id is not None and
945
not _mod_revision.is_null(revision_id)):
946
self.get_revision(revision_id)
948
inter = InterRepository.get(source, self)
950
return inter.fetch(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts)
951
except NotImplementedError:
952
raise errors.IncompatibleRepositories(source, self)
954
def create_bundle(self, target, base, fileobj, format=None):
955
return serializer.write_bundle(self, target, base, fileobj, format)
957
def get_commit_builder(self, branch, parents, config, timestamp=None,
958
timezone=None, committer=None, revprops=None,
960
"""Obtain a CommitBuilder for this repository.
962
:param branch: Branch to commit to.
963
:param parents: Revision ids of the parents of the new revision.
964
:param config: Configuration to use.
965
:param timestamp: Optional timestamp recorded for commit.
966
:param timezone: Optional timezone for timestamp.
967
:param committer: Optional committer to set for commit.
968
:param revprops: Optional dictionary of revision properties.
969
:param revision_id: Optional revision id.
971
result = self._commit_builder_class(self, parents, config,
972
timestamp, timezone, committer, revprops, revision_id)
973
self.start_write_group()
977
if (self.control_files._lock_count == 1 and
978
self.control_files._lock_mode == 'w'):
979
if self._write_group is not None:
980
self.abort_write_group()
981
self.control_files.unlock()
982
raise errors.BzrError(
983
'Must end write groups before releasing write locks.')
984
self.control_files.unlock()
987
def clone(self, a_bzrdir, revision_id=None):
988
"""Clone this repository into a_bzrdir using the current format.
990
Currently no check is made that the format of this repository and
991
the bzrdir format are compatible. FIXME RBC 20060201.
993
:return: The newly created destination repository.
995
# TODO: deprecate after 0.16; cloning this with all its settings is
996
# probably not very useful -- mbp 20070423
997
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
998
self.copy_content_into(dest_repo, revision_id)
1001
def start_write_group(self):
1002
"""Start a write group in the repository.
1004
Write groups are used by repositories which do not have a 1:1 mapping
1005
between file ids and backend store to manage the insertion of data from
1006
both fetch and commit operations.
1008
A write lock is required around the start_write_group/commit_write_group
1009
for the support of lock-requiring repository formats.
1011
One can only insert data into a repository inside a write group.
1015
if not self.is_write_locked():
1016
raise errors.NotWriteLocked(self)
1017
if self._write_group:
1018
raise errors.BzrError('already in a write group')
1019
self._start_write_group()
1020
# so we can detect unlock/relock - the write group is now entered.
1021
self._write_group = self.get_transaction()
1023
def _start_write_group(self):
1024
"""Template method for per-repository write group startup.
1026
This is called before the write group is considered to be
1031
def sprout(self, to_bzrdir, revision_id=None):
1032
"""Create a descendent repository for new development.
1034
Unlike clone, this does not copy the settings of the repository.
1036
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
1037
dest_repo.fetch(self, revision_id=revision_id)
1040
def _create_sprouting_repo(self, a_bzrdir, shared):
1041
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
1042
# use target default format.
1043
dest_repo = a_bzrdir.create_repository()
1045
# Most control formats need the repository to be specifically
1046
# created, but on some old all-in-one formats it's not needed
1048
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
1049
except errors.UninitializableFormat:
1050
dest_repo = a_bzrdir.open_repository()
1054
def has_revision(self, revision_id):
1055
"""True if this repository has a copy of the revision."""
1056
return revision_id in self.has_revisions((revision_id,))
1058
def has_revisions(self, revision_ids):
1059
"""Probe to find out the presence of multiple revisions.
1061
:param revision_ids: An iterable of revision_ids.
1062
:return: A set of the revision_ids that were present.
1064
raise NotImplementedError(self.has_revisions)
1066
return self._revision_store.has_revision_id(revision_id,
1067
self.get_transaction())
1070
def get_revision(self, revision_id):
1071
"""Return the Revision object for a named revision."""
1072
return self.get_revisions([revision_id])[0]
1075
def get_revision_reconcile(self, revision_id):
1076
"""'reconcile' helper routine that allows access to a revision always.
1078
This variant of get_revision does not cross check the weave graph
1079
against the revision one as get_revision does: but it should only
1080
be used by reconcile, or reconcile-alike commands that are correcting
1081
or testing the revision graph.
1083
return self._get_revisions([revision_id])[0]
1086
def get_revisions(self, revision_ids):
1087
"""Get many revisions at once."""
1088
return self._get_revisions(revision_ids)
1091
def _get_revisions(self, revision_ids):
1092
"""Core work logic to get many revisions without sanity checks."""
1093
for rev_id in revision_ids:
1094
if not rev_id or not isinstance(rev_id, basestring):
1095
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1096
revs = self._revision_store.get_revisions(revision_ids,
1097
self.get_transaction())
1101
def get_revision_xml(self, revision_id):
1102
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1103
# would have already do it.
1104
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1105
rev = self.get_revision(revision_id)
1106
rev_tmp = StringIO()
1107
# the current serializer..
1108
self._revision_store._serializer.write_revision(rev, rev_tmp)
1110
return rev_tmp.getvalue()
1113
def get_deltas_for_revisions(self, revisions):
1114
"""Produce a generator of revision deltas.
1116
Note that the input is a sequence of REVISIONS, not revision_ids.
1117
Trees will be held in memory until the generator exits.
1118
Each delta is relative to the revision's lefthand predecessor.
1120
required_trees = set()
1121
for revision in revisions:
1122
required_trees.add(revision.revision_id)
1123
required_trees.update(revision.parent_ids[:1])
1124
trees = dict((t.get_revision_id(), t) for
1125
t in self.revision_trees(required_trees))
1126
for revision in revisions:
1127
if not revision.parent_ids:
1128
old_tree = self.revision_tree(None)
1130
old_tree = trees[revision.parent_ids[0]]
1131
yield trees[revision.revision_id].changes_from(old_tree)
1134
def get_revision_delta(self, revision_id):
1135
"""Return the delta for one revision.
1137
The delta is relative to the left-hand predecessor of the
1140
r = self.get_revision(revision_id)
1141
return list(self.get_deltas_for_revisions([r]))[0]
1144
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1145
signature = gpg_strategy.sign(plaintext)
1146
self.add_signature_text(revision_id, signature)
1149
def add_signature_text(self, revision_id, signature):
1150
self._revision_store.add_revision_signature_text(revision_id,
1152
self.get_transaction())
1154
def find_text_key_references(self):
1155
"""Find the text key references within the repository.
1157
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1158
revision_ids. Each altered file-ids has the exact revision_ids that
1159
altered it listed explicitly.
1160
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1161
to whether they were referred to by the inventory of the
1162
revision_id that they contain. The inventory texts from all present
1163
revision ids are assessed to generate this report.
1165
revision_ids = self.all_revision_ids()
1166
w = self.get_inventory_weave()
1167
pb = ui.ui_factory.nested_progress_bar()
1169
return self._find_text_key_references_from_xml_inventory_lines(
1170
w.iter_lines_added_or_present_in_versions(revision_ids, pb=pb))
1174
def _find_text_key_references_from_xml_inventory_lines(self,
1176
"""Core routine for extracting references to texts from inventories.
1178
This performs the translation of xml lines to revision ids.
1180
:param line_iterator: An iterator of lines, origin_version_id
1181
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1182
to whether they were referred to by the inventory of the
1183
revision_id that they contain. Note that if that revision_id was
1184
not part of the line_iterator's output then False will be given -
1185
even though it may actually refer to that key.
1187
if not self._serializer.support_altered_by_hack:
1188
raise AssertionError(
1189
"_find_text_key_references_from_xml_inventory_lines only "
1190
"supported for branches which store inventory as unnested xml"
1191
", not on %r" % self)
1194
# this code needs to read every new line in every inventory for the
1195
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1196
# not present in one of those inventories is unnecessary but not
1197
# harmful because we are filtering by the revision id marker in the
1198
# inventory lines : we only select file ids altered in one of those
1199
# revisions. We don't need to see all lines in the inventory because
1200
# only those added in an inventory in rev X can contain a revision=X
1202
unescape_revid_cache = {}
1203
unescape_fileid_cache = {}
1205
# jam 20061218 In a big fetch, this handles hundreds of thousands
1206
# of lines, so it has had a lot of inlining and optimizing done.
1207
# Sorry that it is a little bit messy.
1208
# Move several functions to be local variables, since this is a long
1210
search = self._file_ids_altered_regex.search
1211
unescape = _unescape_xml
1212
setdefault = result.setdefault
1213
for line, version_id in line_iterator:
1214
match = search(line)
1217
# One call to match.group() returning multiple items is quite a
1218
# bit faster than 2 calls to match.group() each returning 1
1219
file_id, revision_id = match.group('file_id', 'revision_id')
1221
# Inlining the cache lookups helps a lot when you make 170,000
1222
# lines and 350k ids, versus 8.4 unique ids.
1223
# Using a cache helps in 2 ways:
1224
# 1) Avoids unnecessary decoding calls
1225
# 2) Re-uses cached strings, which helps in future set and
1227
# (2) is enough that removing encoding entirely along with
1228
# the cache (so we are using plain strings) results in no
1229
# performance improvement.
1231
revision_id = unescape_revid_cache[revision_id]
1233
unescaped = unescape(revision_id)
1234
unescape_revid_cache[revision_id] = unescaped
1235
revision_id = unescaped
1237
# Note that unconditionally unescaping means that we deserialise
1238
# every fileid, which for general 'pull' is not great, but we don't
1239
# really want to have some many fulltexts that this matters anyway.
1242
file_id = unescape_fileid_cache[file_id]
1244
unescaped = unescape(file_id)
1245
unescape_fileid_cache[file_id] = unescaped
1248
key = (file_id, revision_id)
1249
setdefault(key, False)
1250
if revision_id == version_id:
1254
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1256
"""Helper routine for fileids_altered_by_revision_ids.
1258
This performs the translation of xml lines to revision ids.
1260
:param line_iterator: An iterator of lines, origin_version_id
1261
:param revision_ids: The revision ids to filter for. This should be a
1262
set or other type which supports efficient __contains__ lookups, as
1263
the revision id from each parsed line will be looked up in the
1264
revision_ids filter.
1265
:return: a dictionary mapping altered file-ids to an iterable of
1266
revision_ids. Each altered file-ids has the exact revision_ids that
1267
altered it listed explicitly.
1270
setdefault = result.setdefault
1271
for file_id, revision_id in \
1272
self._find_text_key_references_from_xml_inventory_lines(
1273
line_iterator).iterkeys():
1274
# once data is all ensured-consistent; then this is
1275
# if revision_id == version_id
1276
if revision_id in revision_ids:
1277
setdefault(file_id, set()).add(revision_id)
1280
def fileids_altered_by_revision_ids(self, revision_ids, _inv_weave=None):
1281
"""Find the file ids and versions affected by revisions.
1283
:param revisions: an iterable containing revision ids.
1284
:param _inv_weave: The inventory weave from this repository or None.
1285
If None, the inventory weave will be opened automatically.
1286
:return: a dictionary mapping altered file-ids to an iterable of
1287
revision_ids. Each altered file-ids has the exact revision_ids that
1288
altered it listed explicitly.
1290
selected_revision_ids = set(revision_ids)
1291
w = _inv_weave or self.get_inventory_weave()
1292
pb = ui.ui_factory.nested_progress_bar()
1294
return self._find_file_ids_from_xml_inventory_lines(
1295
w.iter_lines_added_or_present_in_versions(
1296
selected_revision_ids, pb=pb),
1297
selected_revision_ids)
1301
def iter_files_bytes(self, desired_files):
1302
"""Iterate through file versions.
1304
Files will not necessarily be returned in the order they occur in
1305
desired_files. No specific order is guaranteed.
1307
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1308
value supplied by the caller as part of desired_files. It should
1309
uniquely identify the file version in the caller's context. (Examples:
1310
an index number or a TreeTransform trans_id.)
1312
bytes_iterator is an iterable of bytestrings for the file. The
1313
kind of iterable and length of the bytestrings are unspecified, but for
1314
this implementation, it is a list of lines produced by
1315
VersionedFile.get_lines().
1317
:param desired_files: a list of (file_id, revision_id, identifier)
1320
transaction = self.get_transaction()
1321
for file_id, revision_id, callable_data in desired_files:
1323
weave = self.weave_store.get_weave(file_id, transaction)
1324
except errors.NoSuchFile:
1325
raise errors.NoSuchIdInRepository(self, file_id)
1326
yield callable_data, weave.get_lines(revision_id)
1328
def _generate_text_key_index(self, text_key_references=None,
1330
"""Generate a new text key index for the repository.
1332
This is an expensive function that will take considerable time to run.
1334
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1335
list of parents, also text keys. When a given key has no parents,
1336
the parents list will be [NULL_REVISION].
1338
# All revisions, to find inventory parents.
1339
if ancestors is None:
1340
graph = self.get_graph()
1341
ancestors = graph.get_parent_map(self.all_revision_ids())
1342
if text_key_references is None:
1343
text_key_references = self.find_text_key_references()
1344
pb = ui.ui_factory.nested_progress_bar()
1346
return self._do_generate_text_key_index(ancestors,
1347
text_key_references, pb)
1351
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1352
"""Helper for _generate_text_key_index to avoid deep nesting."""
1353
revision_order = tsort.topo_sort(ancestors)
1354
invalid_keys = set()
1356
for revision_id in revision_order:
1357
revision_keys[revision_id] = set()
1358
text_count = len(text_key_references)
1359
# a cache of the text keys to allow reuse; costs a dict of all the
1360
# keys, but saves a 2-tuple for every child of a given key.
1362
for text_key, valid in text_key_references.iteritems():
1364
invalid_keys.add(text_key)
1366
revision_keys[text_key[1]].add(text_key)
1367
text_key_cache[text_key] = text_key
1368
del text_key_references
1370
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1371
NULL_REVISION = _mod_revision.NULL_REVISION
1372
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1373
# too small for large or very branchy trees. However, for 55K path
1374
# trees, it would be easy to use too much memory trivially. Ideally we
1375
# could gauge this by looking at available real memory etc, but this is
1376
# always a tricky proposition.
1377
inventory_cache = lru_cache.LRUCache(10)
1378
batch_size = 10 # should be ~150MB on a 55K path tree
1379
batch_count = len(revision_order) / batch_size + 1
1381
pb.update("Calculating text parents.", processed_texts, text_count)
1382
for offset in xrange(batch_count):
1383
to_query = revision_order[offset * batch_size:(offset + 1) *
1387
for rev_tree in self.revision_trees(to_query):
1388
revision_id = rev_tree.get_revision_id()
1389
parent_ids = ancestors[revision_id]
1390
for text_key in revision_keys[revision_id]:
1391
pb.update("Calculating text parents.", processed_texts)
1392
processed_texts += 1
1393
candidate_parents = []
1394
for parent_id in parent_ids:
1395
parent_text_key = (text_key[0], parent_id)
1397
check_parent = parent_text_key not in \
1398
revision_keys[parent_id]
1400
# the parent parent_id is a ghost:
1401
check_parent = False
1402
# truncate the derived graph against this ghost.
1403
parent_text_key = None
1405
# look at the parent commit details inventories to
1406
# determine possible candidates in the per file graph.
1409
inv = inventory_cache[parent_id]
1411
inv = self.revision_tree(parent_id).inventory
1412
inventory_cache[parent_id] = inv
1413
parent_entry = inv._byid.get(text_key[0], None)
1414
if parent_entry is not None:
1416
text_key[0], parent_entry.revision)
1418
parent_text_key = None
1419
if parent_text_key is not None:
1420
candidate_parents.append(
1421
text_key_cache[parent_text_key])
1422
parent_heads = text_graph.heads(candidate_parents)
1423
new_parents = list(parent_heads)
1424
new_parents.sort(key=lambda x:candidate_parents.index(x))
1425
if new_parents == []:
1426
new_parents = [NULL_REVISION]
1427
text_index[text_key] = new_parents
1429
for text_key in invalid_keys:
1430
text_index[text_key] = [NULL_REVISION]
1433
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1434
"""Get an iterable listing the keys of all the data introduced by a set
1437
The keys will be ordered so that the corresponding items can be safely
1438
fetched and inserted in that order.
1440
:returns: An iterable producing tuples of (knit-kind, file-id,
1441
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1442
'revisions'. file-id is None unless knit-kind is 'file'.
1444
# XXX: it's a bit weird to control the inventory weave caching in this
1445
# generator. Ideally the caching would be done in fetch.py I think. Or
1446
# maybe this generator should explicitly have the contract that it
1447
# should not be iterated until the previously yielded item has been
1450
inv_w = self.get_inventory_weave()
1452
# file ids that changed
1453
file_ids = self.fileids_altered_by_revision_ids(revision_ids, inv_w)
1455
num_file_ids = len(file_ids)
1456
for file_id, altered_versions in file_ids.iteritems():
1457
if _files_pb is not None:
1458
_files_pb.update("fetch texts", count, num_file_ids)
1460
yield ("file", file_id, altered_versions)
1461
# We're done with the files_pb. Note that it finished by the caller,
1462
# just as it was created by the caller.
1466
yield ("inventory", None, revision_ids)
1469
revisions_with_signatures = set()
1470
for rev_id in revision_ids:
1472
self.get_signature_text(rev_id)
1473
except errors.NoSuchRevision:
1477
revisions_with_signatures.add(rev_id)
1479
yield ("signatures", None, revisions_with_signatures)
1482
yield ("revisions", None, revision_ids)
1485
def get_inventory_weave(self):
1486
return self.control_weaves.get_weave('inventory',
1487
self.get_transaction())
1490
def get_inventory(self, revision_id):
1491
"""Get Inventory object by revision id."""
1492
return self.iter_inventories([revision_id]).next()
1494
def iter_inventories(self, revision_ids):
1495
"""Get many inventories by revision_ids.
1497
This will buffer some or all of the texts used in constructing the
1498
inventories in memory, but will only parse a single inventory at a
1501
:return: An iterator of inventories.
1503
if ((None in revision_ids)
1504
or (_mod_revision.NULL_REVISION in revision_ids)):
1505
raise ValueError('cannot get null revision inventory')
1506
return self._iter_inventories(revision_ids)
1508
def _iter_inventories(self, revision_ids):
1509
"""single-document based inventory iteration."""
1510
texts = self.get_inventory_weave().get_texts(revision_ids)
1511
for text, revision_id in zip(texts, revision_ids):
1512
yield self.deserialise_inventory(revision_id, text)
1514
def deserialise_inventory(self, revision_id, xml):
1515
"""Transform the xml into an inventory object.
1517
:param revision_id: The expected revision id of the inventory.
1518
:param xml: A serialised inventory.
1520
result = self._serializer.read_inventory_from_string(xml, revision_id)
1521
if result.revision_id != revision_id:
1522
raise AssertionError('revision id mismatch %s != %s' % (
1523
result.revision_id, revision_id))
1526
def serialise_inventory(self, inv):
1527
return self._serializer.write_inventory_to_string(inv)
1529
def _serialise_inventory_to_lines(self, inv):
1530
return self._serializer.write_inventory_to_lines(inv)
1532
def get_serializer_format(self):
1533
return self._serializer.format_num
1536
def get_inventory_xml(self, revision_id):
1537
"""Get inventory XML as a file object."""
1539
iw = self.get_inventory_weave()
1540
return iw.get_text(revision_id)
1542
raise errors.HistoryMissing(self, 'inventory', revision_id)
1545
def get_inventory_sha1(self, revision_id):
1546
"""Return the sha1 hash of the inventory entry
1548
return self.get_revision(revision_id).inventory_sha1
1551
@deprecated_method(symbol_versioning.one_four)
1552
def get_revision_graph(self, revision_id=None):
1553
"""Return a dictionary containing the revision graph.
1555
NB: This method should not be used as it accesses the entire graph all
1556
at once, which is much more data than most operations should require.
1558
:param revision_id: The revision_id to get a graph from. If None, then
1559
the entire revision graph is returned. This is a deprecated mode of
1560
operation and will be removed in the future.
1561
:return: a dictionary of revision_id->revision_parents_list.
1563
raise NotImplementedError(self.get_revision_graph)
1566
@deprecated_method(symbol_versioning.one_three)
1567
def get_revision_graph_with_ghosts(self, revision_ids=None):
1568
"""Return a graph of the revisions with ghosts marked as applicable.
1570
:param revision_ids: an iterable of revisions to graph or None for all.
1571
:return: a Graph object with the graph reachable from revision_ids.
1573
if 'evil' in debug.debug_flags:
1575
"get_revision_graph_with_ghosts scales with size of history.")
1576
result = deprecated_graph.Graph()
1577
if not revision_ids:
1578
pending = set(self.all_revision_ids())
1581
pending = set(revision_ids)
1582
# special case NULL_REVISION
1583
if _mod_revision.NULL_REVISION in pending:
1584
pending.remove(_mod_revision.NULL_REVISION)
1585
required = set(pending)
1588
revision_id = pending.pop()
1590
rev = self.get_revision(revision_id)
1591
except errors.NoSuchRevision:
1592
if revision_id in required:
1595
result.add_ghost(revision_id)
1597
for parent_id in rev.parent_ids:
1598
# is this queued or done ?
1599
if (parent_id not in pending and
1600
parent_id not in done):
1602
pending.add(parent_id)
1603
result.add_node(revision_id, rev.parent_ids)
1604
done.add(revision_id)
1607
def iter_reverse_revision_history(self, revision_id):
1608
"""Iterate backwards through revision ids in the lefthand history
1610
:param revision_id: The revision id to start with. All its lefthand
1611
ancestors will be traversed.
1613
graph = self.get_graph()
1614
next_id = revision_id
1616
if next_id in (None, _mod_revision.NULL_REVISION):
1619
# Note: The following line may raise KeyError in the event of
1620
# truncated history. We decided not to have a try:except:raise
1621
# RevisionNotPresent here until we see a use for it, because of the
1622
# cost in an inner loop that is by its very nature O(history).
1623
# Robert Collins 20080326
1624
parents = graph.get_parent_map([next_id])[next_id]
1625
if len(parents) == 0:
1628
next_id = parents[0]
1631
def get_revision_inventory(self, revision_id):
1632
"""Return inventory of a past revision."""
1633
# TODO: Unify this with get_inventory()
1634
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1635
# must be the same as its revision, so this is trivial.
1636
if revision_id is None:
1637
# This does not make sense: if there is no revision,
1638
# then it is the current tree inventory surely ?!
1639
# and thus get_root_id() is something that looks at the last
1640
# commit on the branch, and the get_root_id is an inventory check.
1641
raise NotImplementedError
1642
# return Inventory(self.get_root_id())
1644
return self.get_inventory(revision_id)
1647
def is_shared(self):
1648
"""Return True if this repository is flagged as a shared repository."""
1649
raise NotImplementedError(self.is_shared)
1652
def reconcile(self, other=None, thorough=False):
1653
"""Reconcile this repository."""
1654
from bzrlib.reconcile import RepoReconciler
1655
reconciler = RepoReconciler(self, thorough=thorough)
1656
reconciler.reconcile()
1659
def _refresh_data(self):
1660
"""Helper called from lock_* to ensure coherency with disk.
1662
The default implementation does nothing; it is however possible
1663
for repositories to maintain loaded indices across multiple locks
1664
by checking inside their implementation of this method to see
1665
whether their indices are still valid. This depends of course on
1666
the disk format being validatable in this manner.
1670
def revision_tree(self, revision_id):
1671
"""Return Tree for a revision on this branch.
1673
`revision_id` may be None for the empty tree revision.
1675
# TODO: refactor this to use an existing revision object
1676
# so we don't need to read it in twice.
1677
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1678
return RevisionTree(self, Inventory(root_id=None),
1679
_mod_revision.NULL_REVISION)
1681
inv = self.get_revision_inventory(revision_id)
1682
return RevisionTree(self, inv, revision_id)
1685
def revision_trees(self, revision_ids):
1686
"""Return Tree for a revision on this branch.
1688
`revision_id` may not be None or 'null:'"""
1689
inventories = self.iter_inventories(revision_ids)
1690
for inv in inventories:
1691
yield RevisionTree(self, inv, inv.revision_id)
1694
def get_ancestry(self, revision_id, topo_sorted=True):
1695
"""Return a list of revision-ids integrated by a revision.
1697
The first element of the list is always None, indicating the origin
1698
revision. This might change when we have history horizons, or
1699
perhaps we should have a new API.
1701
This is topologically sorted.
1703
if _mod_revision.is_null(revision_id):
1705
if not self.has_revision(revision_id):
1706
raise errors.NoSuchRevision(self, revision_id)
1707
w = self.get_inventory_weave()
1708
candidates = w.get_ancestry(revision_id, topo_sorted)
1709
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1712
"""Compress the data within the repository.
1714
This operation only makes sense for some repository types. For other
1715
types it should be a no-op that just returns.
1717
This stub method does not require a lock, but subclasses should use
1718
@needs_write_lock as this is a long running call its reasonable to
1719
implicitly lock for the user.
1723
def print_file(self, file, revision_id):
1724
"""Print `file` to stdout.
1726
FIXME RBC 20060125 as John Meinel points out this is a bad api
1727
- it writes to stdout, it assumes that that is valid etc. Fix
1728
by creating a new more flexible convenience function.
1730
tree = self.revision_tree(revision_id)
1731
# use inventory as it was in that revision
1732
file_id = tree.inventory.path2id(file)
1734
# TODO: jam 20060427 Write a test for this code path
1735
# it had a bug in it, and was raising the wrong
1737
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1738
tree.print_file(file_id)
1740
def get_transaction(self):
1741
return self.control_files.get_transaction()
1743
def revision_parents(self, revision_id):
1744
return self.get_inventory_weave().parent_names(revision_id)
1746
@deprecated_method(symbol_versioning.one_one)
1747
def get_parents(self, revision_ids):
1748
"""See StackedParentsProvider.get_parents"""
1749
parent_map = self.get_parent_map(revision_ids)
1750
return [parent_map.get(r, None) for r in revision_ids]
1752
def get_parent_map(self, keys):
1753
"""See graph._StackedParentsProvider.get_parent_map"""
1755
for revision_id in keys:
1756
if revision_id == _mod_revision.NULL_REVISION:
1757
parent_map[revision_id] = ()
1760
parent_id_list = self.get_revision(revision_id).parent_ids
1761
except errors.NoSuchRevision:
1764
if len(parent_id_list) == 0:
1765
parent_ids = (_mod_revision.NULL_REVISION,)
1767
parent_ids = tuple(parent_id_list)
1768
parent_map[revision_id] = parent_ids
1771
def _make_parents_provider(self):
1774
def get_graph(self, other_repository=None):
1775
"""Return the graph walker for this repository format"""
1776
parents_provider = self._make_parents_provider()
1777
if (other_repository is not None and
1778
not self.has_same_location(other_repository)):
1779
parents_provider = graph._StackedParentsProvider(
1780
[parents_provider, other_repository._make_parents_provider()])
1781
return graph.Graph(parents_provider)
1783
def _get_versioned_file_checker(self):
1784
"""Return an object suitable for checking versioned files."""
1785
return _VersionedFileChecker(self)
1787
def revision_ids_to_search_result(self, result_set):
1788
"""Convert a set of revision ids to a graph SearchResult."""
1789
result_parents = set()
1790
for parents in self.get_graph().get_parent_map(
1791
result_set).itervalues():
1792
result_parents.update(parents)
1793
included_keys = result_set.intersection(result_parents)
1794
start_keys = result_set.difference(included_keys)
1795
exclude_keys = result_parents.difference(result_set)
1796
result = graph.SearchResult(start_keys, exclude_keys,
1797
len(result_set), result_set)
1801
def set_make_working_trees(self, new_value):
1802
"""Set the policy flag for making working trees when creating branches.
1804
This only applies to branches that use this repository.
1806
The default is 'True'.
1807
:param new_value: True to restore the default, False to disable making
1810
raise NotImplementedError(self.set_make_working_trees)
1812
def make_working_trees(self):
1813
"""Returns the policy for making working trees on new branches."""
1814
raise NotImplementedError(self.make_working_trees)
1817
def sign_revision(self, revision_id, gpg_strategy):
1818
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1819
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1822
def has_signature_for_revision_id(self, revision_id):
1823
"""Query for a revision signature for revision_id in the repository."""
1824
return self._revision_store.has_signature(revision_id,
1825
self.get_transaction())
1828
def get_signature_text(self, revision_id):
1829
"""Return the text for a signature."""
1830
return self._revision_store.get_signature_text(revision_id,
1831
self.get_transaction())
1834
def check(self, revision_ids=None):
1835
"""Check consistency of all history of given revision_ids.
1837
Different repository implementations should override _check().
1839
:param revision_ids: A non-empty list of revision_ids whose ancestry
1840
will be checked. Typically the last revision_id of a branch.
1842
return self._check(revision_ids)
1844
def _check(self, revision_ids):
1845
result = check.Check(self)
1849
def _warn_if_deprecated(self):
1850
global _deprecation_warning_done
1851
if _deprecation_warning_done:
1853
_deprecation_warning_done = True
1854
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1855
% (self._format, self.bzrdir.transport.base))
1857
def supports_rich_root(self):
1858
return self._format.rich_root_data
1860
def _check_ascii_revisionid(self, revision_id, method):
1861
"""Private helper for ascii-only repositories."""
1862
# weave repositories refuse to store revisionids that are non-ascii.
1863
if revision_id is not None:
1864
# weaves require ascii revision ids.
1865
if isinstance(revision_id, unicode):
1867
revision_id.encode('ascii')
1868
except UnicodeEncodeError:
1869
raise errors.NonAsciiRevisionId(method, self)
1872
revision_id.decode('ascii')
1873
except UnicodeDecodeError:
1874
raise errors.NonAsciiRevisionId(method, self)
1876
def revision_graph_can_have_wrong_parents(self):
1877
"""Is it possible for this repository to have a revision graph with
1880
If True, then this repository must also implement
1881
_find_inconsistent_revision_parents so that check and reconcile can
1882
check for inconsistencies before proceeding with other checks that may
1883
depend on the revision index being consistent.
1885
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1888
# remove these delegates a while after bzr 0.15
1889
def __make_delegated(name, from_module):
1890
def _deprecated_repository_forwarder():
1891
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1892
% (name, from_module),
1895
m = __import__(from_module, globals(), locals(), [name])
1897
return getattr(m, name)
1898
except AttributeError:
1899
raise AttributeError('module %s has no name %s'
1901
globals()[name] = _deprecated_repository_forwarder
1904
'AllInOneRepository',
1905
'WeaveMetaDirRepository',
1906
'PreSplitOutRepositoryFormat',
1907
'RepositoryFormat4',
1908
'RepositoryFormat5',
1909
'RepositoryFormat6',
1910
'RepositoryFormat7',
1912
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1916
'RepositoryFormatKnit',
1917
'RepositoryFormatKnit1',
1919
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1922
def install_revision(repository, rev, revision_tree):
1923
"""Install all revision data into a repository."""
1924
install_revisions(repository, [(rev, revision_tree, None)])
1927
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1928
"""Install all revision data into a repository.
1930
Accepts an iterable of revision, tree, signature tuples. The signature
1933
repository.start_write_group()
1935
for n, (revision, revision_tree, signature) in enumerate(iterable):
1936
_install_revision(repository, revision, revision_tree, signature)
1938
pb.update('Transferring revisions', n + 1, num_revisions)
1940
repository.abort_write_group()
1943
repository.commit_write_group()
1946
def _install_revision(repository, rev, revision_tree, signature):
1947
"""Install all revision data into a repository."""
1948
present_parents = []
1950
for p_id in rev.parent_ids:
1951
if repository.has_revision(p_id):
1952
present_parents.append(p_id)
1953
parent_trees[p_id] = repository.revision_tree(p_id)
1955
parent_trees[p_id] = repository.revision_tree(None)
1957
inv = revision_tree.inventory
1958
entries = inv.iter_entries()
1959
# backwards compatibility hack: skip the root id.
1960
if not repository.supports_rich_root():
1961
path, root = entries.next()
1962
if root.revision != rev.revision_id:
1963
raise errors.IncompatibleRevision(repr(repository))
1964
# Add the texts that are not already present
1965
for path, ie in entries:
1966
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1967
repository.get_transaction())
1968
if ie.revision not in w:
1970
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1971
# with InventoryEntry.find_previous_heads(). if it is, then there
1972
# is a latent bug here where the parents may have ancestors of each
1974
for revision, tree in parent_trees.iteritems():
1975
if ie.file_id not in tree:
1977
parent_id = tree.inventory[ie.file_id].revision
1978
if parent_id in text_parents:
1980
text_parents.append(parent_id)
1982
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1983
repository.get_transaction())
1984
lines = revision_tree.get_file(ie.file_id).readlines()
1985
vfile.add_lines(rev.revision_id, text_parents, lines)
1987
# install the inventory
1988
repository.add_inventory(rev.revision_id, inv, present_parents)
1989
except errors.RevisionAlreadyPresent:
1991
if signature is not None:
1992
repository.add_signature_text(rev.revision_id, signature)
1993
repository.add_revision(rev.revision_id, rev, inv)
1996
class MetaDirRepository(Repository):
1997
"""Repositories in the new meta-dir layout."""
1999
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
2000
super(MetaDirRepository, self).__init__(_format,
2006
dir_mode = self.control_files._dir_mode
2007
file_mode = self.control_files._file_mode
2010
def is_shared(self):
2011
"""Return True if this repository is flagged as a shared repository."""
2012
return self.control_files._transport.has('shared-storage')
2015
def set_make_working_trees(self, new_value):
2016
"""Set the policy flag for making working trees when creating branches.
2018
This only applies to branches that use this repository.
2020
The default is 'True'.
2021
:param new_value: True to restore the default, False to disable making
2026
self.control_files._transport.delete('no-working-trees')
2027
except errors.NoSuchFile:
2030
self.control_files.put_utf8('no-working-trees', '')
2032
def make_working_trees(self):
2033
"""Returns the policy for making working trees on new branches."""
2034
return not self.control_files._transport.has('no-working-trees')
2037
class MetaDirVersionedFileRepository(MetaDirRepository):
2038
"""Repositories in a meta-dir, that work via versioned file objects."""
2040
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
2041
super(MetaDirVersionedFileRepository, self).__init__(_format, a_bzrdir,
2042
control_files, _revision_store, control_store, text_store)
2043
_revision_store.get_scope = self.get_transaction
2044
control_store.get_scope = self.get_transaction
2045
text_store.get_scope = self.get_transaction
2048
class RepositoryFormatRegistry(registry.Registry):
2049
"""Registry of RepositoryFormats."""
2051
def get(self, format_string):
2052
r = registry.Registry.get(self, format_string)
2058
format_registry = RepositoryFormatRegistry()
2059
"""Registry of formats, indexed by their identifying format string.
2061
This can contain either format instances themselves, or classes/factories that
2062
can be called to obtain one.
2066
#####################################################################
2067
# Repository Formats
2069
class RepositoryFormat(object):
2070
"""A repository format.
2072
Formats provide three things:
2073
* An initialization routine to construct repository data on disk.
2074
* a format string which is used when the BzrDir supports versioned
2076
* an open routine which returns a Repository instance.
2078
There is one and only one Format subclass for each on-disk format. But
2079
there can be one Repository subclass that is used for several different
2080
formats. The _format attribute on a Repository instance can be used to
2081
determine the disk format.
2083
Formats are placed in an dict by their format string for reference
2084
during opening. These should be subclasses of RepositoryFormat
2087
Once a format is deprecated, just deprecate the initialize and open
2088
methods on the format class. Do not deprecate the object, as the
2089
object will be created every system load.
2091
Common instance attributes:
2092
_matchingbzrdir - the bzrdir format that the repository format was
2093
originally written to work with. This can be used if manually
2094
constructing a bzrdir and repository, or more commonly for test suite
2098
# Set to True or False in derived classes. True indicates that the format
2099
# supports ghosts gracefully.
2100
supports_ghosts = None
2101
# Can this repository be given external locations to lookup additional
2102
# data. Set to True or False in derived classes.
2103
supports_external_lookups = None
2106
return "<%s>" % self.__class__.__name__
2108
def __eq__(self, other):
2109
# format objects are generally stateless
2110
return isinstance(other, self.__class__)
2112
def __ne__(self, other):
2113
return not self == other
2116
def find_format(klass, a_bzrdir):
2117
"""Return the format for the repository object in a_bzrdir.
2119
This is used by bzr native formats that have a "format" file in
2120
the repository. Other methods may be used by different types of
2124
transport = a_bzrdir.get_repository_transport(None)
2125
format_string = transport.get("format").read()
2126
return format_registry.get(format_string)
2127
except errors.NoSuchFile:
2128
raise errors.NoRepositoryPresent(a_bzrdir)
2130
raise errors.UnknownFormatError(format=format_string,
2134
def register_format(klass, format):
2135
format_registry.register(format.get_format_string(), format)
2138
def unregister_format(klass, format):
2139
format_registry.remove(format.get_format_string())
2142
def get_default_format(klass):
2143
"""Return the current default format."""
2144
from bzrlib import bzrdir
2145
return bzrdir.format_registry.make_bzrdir('default').repository_format
2147
def _get_control_store(self, repo_transport, control_files):
2148
"""Return the control store for this repository."""
2149
raise NotImplementedError(self._get_control_store)
2151
def get_format_string(self):
2152
"""Return the ASCII format string that identifies this format.
2154
Note that in pre format ?? repositories the format string is
2155
not permitted nor written to disk.
2157
raise NotImplementedError(self.get_format_string)
2159
def get_format_description(self):
2160
"""Return the short description for this format."""
2161
raise NotImplementedError(self.get_format_description)
2163
def _get_revision_store(self, repo_transport, control_files):
2164
"""Return the revision store object for this a_bzrdir."""
2165
raise NotImplementedError(self._get_revision_store)
2167
def _get_text_rev_store(self,
2174
"""Common logic for getting a revision store for a repository.
2176
see self._get_revision_store for the subclass-overridable method to
2177
get the store for a repository.
2179
from bzrlib.store.revision.text import TextRevisionStore
2180
dir_mode = control_files._dir_mode
2181
file_mode = control_files._file_mode
2182
text_store = TextStore(transport.clone(name),
2184
compressed=compressed,
2186
file_mode=file_mode)
2187
_revision_store = TextRevisionStore(text_store, serializer)
2188
return _revision_store
2190
# TODO: this shouldn't be in the base class, it's specific to things that
2191
# use weaves or knits -- mbp 20070207
2192
def _get_versioned_file_store(self,
2197
versionedfile_class=None,
2198
versionedfile_kwargs={},
2200
if versionedfile_class is None:
2201
versionedfile_class = self._versionedfile_class
2202
weave_transport = control_files._transport.clone(name)
2203
dir_mode = control_files._dir_mode
2204
file_mode = control_files._file_mode
2205
return VersionedFileStore(weave_transport, prefixed=prefixed,
2207
file_mode=file_mode,
2208
versionedfile_class=versionedfile_class,
2209
versionedfile_kwargs=versionedfile_kwargs,
2212
def initialize(self, a_bzrdir, shared=False):
2213
"""Initialize a repository of this format in a_bzrdir.
2215
:param a_bzrdir: The bzrdir to put the new repository in it.
2216
:param shared: The repository should be initialized as a sharable one.
2217
:returns: The new repository object.
2219
This may raise UninitializableFormat if shared repository are not
2220
compatible the a_bzrdir.
2222
raise NotImplementedError(self.initialize)
2224
def is_supported(self):
2225
"""Is this format supported?
2227
Supported formats must be initializable and openable.
2228
Unsupported formats may not support initialization or committing or
2229
some other features depending on the reason for not being supported.
2233
def check_conversion_target(self, target_format):
2234
raise NotImplementedError(self.check_conversion_target)
2236
def open(self, a_bzrdir, _found=False):
2237
"""Return an instance of this format for the bzrdir a_bzrdir.
2239
_found is a private parameter, do not use it.
2241
raise NotImplementedError(self.open)
2244
class MetaDirRepositoryFormat(RepositoryFormat):
2245
"""Common base class for the new repositories using the metadir layout."""
2247
rich_root_data = False
2248
supports_tree_reference = False
2249
supports_external_lookups = False
2250
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
2253
super(MetaDirRepositoryFormat, self).__init__()
2255
def _create_control_files(self, a_bzrdir):
2256
"""Create the required files and the initial control_files object."""
2257
# FIXME: RBC 20060125 don't peek under the covers
2258
# NB: no need to escape relative paths that are url safe.
2259
repository_transport = a_bzrdir.get_repository_transport(self)
2260
control_files = lockable_files.LockableFiles(repository_transport,
2261
'lock', lockdir.LockDir)
2262
control_files.create_lock()
2263
return control_files
2265
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
2266
"""Upload the initial blank content."""
2267
control_files = self._create_control_files(a_bzrdir)
2268
control_files.lock_write()
2270
control_files._transport.mkdir_multi(dirs,
2271
mode=control_files._dir_mode)
2272
for file, content in files:
2273
control_files.put(file, content)
2274
for file, content in utf8_files:
2275
control_files.put_utf8(file, content)
2277
control_files.put_utf8('shared-storage', '')
2279
control_files.unlock()
2282
# formats which have no format string are not discoverable
2283
# and not independently creatable, so are not registered. They're
2284
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
2285
# needed, it's constructed directly by the BzrDir. Non-native formats where
2286
# the repository is not separately opened are similar.
2288
format_registry.register_lazy(
2289
'Bazaar-NG Repository format 7',
2290
'bzrlib.repofmt.weaverepo',
2294
format_registry.register_lazy(
2295
'Bazaar-NG Knit Repository Format 1',
2296
'bzrlib.repofmt.knitrepo',
2297
'RepositoryFormatKnit1',
2300
format_registry.register_lazy(
2301
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
2302
'bzrlib.repofmt.knitrepo',
2303
'RepositoryFormatKnit3',
2306
format_registry.register_lazy(
2307
'Bazaar Knit Repository Format 4 (bzr 1.0)\n',
2308
'bzrlib.repofmt.knitrepo',
2309
'RepositoryFormatKnit4',
2312
# Pack-based formats. There is one format for pre-subtrees, and one for
2313
# post-subtrees to allow ease of testing.
2314
# NOTE: These are experimental in 0.92. Stable in 1.0 and above
2315
format_registry.register_lazy(
2316
'Bazaar pack repository format 1 (needs bzr 0.92)\n',
2317
'bzrlib.repofmt.pack_repo',
2318
'RepositoryFormatKnitPack1',
2320
format_registry.register_lazy(
2321
'Bazaar pack repository format 1 with subtree support (needs bzr 0.92)\n',
2322
'bzrlib.repofmt.pack_repo',
2323
'RepositoryFormatKnitPack3',
2325
format_registry.register_lazy(
2326
'Bazaar pack repository format 1 with rich root (needs bzr 1.0)\n',
2327
'bzrlib.repofmt.pack_repo',
2328
'RepositoryFormatKnitPack4',
2330
# Development formats.
2332
# development 0 - stub to introduce development versioning scheme.
2333
format_registry.register_lazy(
2334
"Bazaar development format 0 (needs bzr.dev from before 1.3)\n",
2335
'bzrlib.repofmt.pack_repo',
2336
'RepositoryFormatPackDevelopment0',
2338
format_registry.register_lazy(
2339
("Bazaar development format 0 with subtree support "
2340
"(needs bzr.dev from before 1.3)\n"),
2341
'bzrlib.repofmt.pack_repo',
2342
'RepositoryFormatPackDevelopment0Subtree',
2344
# 1.3->1.4 go below here
2347
class InterRepository(InterObject):
2348
"""This class represents operations taking place between two repositories.
2350
Its instances have methods like copy_content and fetch, and contain
2351
references to the source and target repositories these operations can be
2354
Often we will provide convenience methods on 'repository' which carry out
2355
operations with another repository - they will always forward to
2356
InterRepository.get(other).method_name(parameters).
2360
"""The available optimised InterRepository types."""
2362
def copy_content(self, revision_id=None):
2363
raise NotImplementedError(self.copy_content)
2365
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2366
"""Fetch the content required to construct revision_id.
2368
The content is copied from self.source to self.target.
2370
:param revision_id: if None all content is copied, if NULL_REVISION no
2372
:param pb: optional progress bar to use for progress reports. If not
2373
provided a default one will be created.
2375
Returns the copied revision count and the failed revisions in a tuple:
2378
raise NotImplementedError(self.fetch)
2380
def _walk_to_common_revisions(self, revision_ids):
2381
"""Walk out from revision_ids in source to revisions target has.
2383
:param revision_ids: The start point for the search.
2384
:return: A set of revision ids.
2386
graph = self.source.get_graph()
2387
target_graph = self.target.get_graph()
2388
missing_revs = set()
2389
# ensure we don't pay silly lookup costs.
2390
revision_ids = frozenset(revision_ids)
2391
searcher = graph._make_breadth_first_searcher(revision_ids)
2392
null_set = frozenset([_mod_revision.NULL_REVISION])
2395
next_revs, ghosts = searcher.next_with_ghosts()
2396
except StopIteration:
2398
if revision_ids.intersection(ghosts):
2399
absent_ids = set(revision_ids.intersection(ghosts))
2400
# If all absent_ids are present in target, no error is needed.
2401
absent_ids.difference_update(
2402
set(target_graph.get_parent_map(absent_ids)))
2404
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2405
# we don't care about other ghosts as we can't fetch them and
2406
# haven't been asked to.
2407
next_revs = set(next_revs)
2408
# we always have NULL_REVISION present.
2409
have_revs = set(target_graph.get_parent_map(next_revs)).union(null_set)
2410
missing_revs.update(next_revs - have_revs)
2411
searcher.stop_searching_any(have_revs)
2412
return searcher.get_result()
2414
@deprecated_method(symbol_versioning.one_two)
2416
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2417
"""Return the revision ids that source has that target does not.
2419
These are returned in topological order.
2421
:param revision_id: only return revision ids included by this
2423
:param find_ghosts: If True find missing revisions in deep history
2424
rather than just finding the surface difference.
2426
return list(self.search_missing_revision_ids(
2427
revision_id, find_ghosts).get_keys())
2430
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2431
"""Return the revision ids that source has that target does not.
2433
:param revision_id: only return revision ids included by this
2435
:param find_ghosts: If True find missing revisions in deep history
2436
rather than just finding the surface difference.
2437
:return: A bzrlib.graph.SearchResult.
2439
# stop searching at found target revisions.
2440
if not find_ghosts and revision_id is not None:
2441
return self._walk_to_common_revisions([revision_id])
2442
# generic, possibly worst case, slow code path.
2443
target_ids = set(self.target.all_revision_ids())
2444
if revision_id is not None:
2445
source_ids = self.source.get_ancestry(revision_id)
2446
if source_ids[0] is not None:
2447
raise AssertionError()
2450
source_ids = self.source.all_revision_ids()
2451
result_set = set(source_ids).difference(target_ids)
2452
return self.source.revision_ids_to_search_result(result_set)
2455
def _same_model(source, target):
2456
"""True if source and target have the same data representation."""
2457
if source.supports_rich_root() != target.supports_rich_root():
2459
if source._serializer != target._serializer:
2464
class InterSameDataRepository(InterRepository):
2465
"""Code for converting between repositories that represent the same data.
2467
Data format and model must match for this to work.
2471
def _get_repo_format_to_test(self):
2472
"""Repository format for testing with.
2474
InterSameData can pull from subtree to subtree and from non-subtree to
2475
non-subtree, so we test this with the richest repository format.
2477
from bzrlib.repofmt import knitrepo
2478
return knitrepo.RepositoryFormatKnit3()
2481
def is_compatible(source, target):
2482
return InterRepository._same_model(source, target)
2485
def copy_content(self, revision_id=None):
2486
"""Make a complete copy of the content in self into destination.
2488
This copies both the repository's revision data, and configuration information
2489
such as the make_working_trees setting.
2491
This is a destructive operation! Do not use it on existing
2494
:param revision_id: Only copy the content needed to construct
2495
revision_id and its parents.
2498
self.target.set_make_working_trees(self.source.make_working_trees())
2499
except NotImplementedError:
2501
# but don't bother fetching if we have the needed data now.
2502
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2503
self.target.has_revision(revision_id)):
2505
self.target.fetch(self.source, revision_id=revision_id)
2508
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2509
"""See InterRepository.fetch()."""
2510
from bzrlib.fetch import GenericRepoFetcher
2511
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2512
self.source, self.source._format, self.target,
2513
self.target._format)
2514
f = GenericRepoFetcher(to_repository=self.target,
2515
from_repository=self.source,
2516
last_revision=revision_id,
2517
pb=pb, find_ghosts=find_ghosts)
2518
return f.count_copied, f.failed_revisions
2521
class InterWeaveRepo(InterSameDataRepository):
2522
"""Optimised code paths between Weave based repositories.
2524
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2525
implemented lazy inter-object optimisation.
2529
def _get_repo_format_to_test(self):
2530
from bzrlib.repofmt import weaverepo
2531
return weaverepo.RepositoryFormat7()
2534
def is_compatible(source, target):
2535
"""Be compatible with known Weave formats.
2537
We don't test for the stores being of specific types because that
2538
could lead to confusing results, and there is no need to be
2541
from bzrlib.repofmt.weaverepo import (
2547
return (isinstance(source._format, (RepositoryFormat5,
2549
RepositoryFormat7)) and
2550
isinstance(target._format, (RepositoryFormat5,
2552
RepositoryFormat7)))
2553
except AttributeError:
2557
def copy_content(self, revision_id=None):
2558
"""See InterRepository.copy_content()."""
2559
# weave specific optimised path:
2561
self.target.set_make_working_trees(self.source.make_working_trees())
2562
except (errors.RepositoryUpgradeRequired, NotImplemented):
2564
# FIXME do not peek!
2565
if self.source.control_files._transport.listable():
2566
pb = ui.ui_factory.nested_progress_bar()
2568
self.target.weave_store.copy_all_ids(
2569
self.source.weave_store,
2571
from_transaction=self.source.get_transaction(),
2572
to_transaction=self.target.get_transaction())
2573
pb.update('copying inventory', 0, 1)
2574
self.target.control_weaves.copy_multi(
2575
self.source.control_weaves, ['inventory'],
2576
from_transaction=self.source.get_transaction(),
2577
to_transaction=self.target.get_transaction())
2578
self.target._revision_store.text_store.copy_all_ids(
2579
self.source._revision_store.text_store,
2584
self.target.fetch(self.source, revision_id=revision_id)
2587
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2588
"""See InterRepository.fetch()."""
2589
from bzrlib.fetch import GenericRepoFetcher
2590
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2591
self.source, self.source._format, self.target, self.target._format)
2592
f = GenericRepoFetcher(to_repository=self.target,
2593
from_repository=self.source,
2594
last_revision=revision_id,
2595
pb=pb, find_ghosts=find_ghosts)
2596
return f.count_copied, f.failed_revisions
2599
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2600
"""See InterRepository.missing_revision_ids()."""
2601
# we want all revisions to satisfy revision_id in source.
2602
# but we don't want to stat every file here and there.
2603
# we want then, all revisions other needs to satisfy revision_id
2604
# checked, but not those that we have locally.
2605
# so the first thing is to get a subset of the revisions to
2606
# satisfy revision_id in source, and then eliminate those that
2607
# we do already have.
2608
# this is slow on high latency connection to self, but as as this
2609
# disk format scales terribly for push anyway due to rewriting
2610
# inventory.weave, this is considered acceptable.
2612
if revision_id is not None:
2613
source_ids = self.source.get_ancestry(revision_id)
2614
if source_ids[0] is not None:
2615
raise AssertionError()
2618
source_ids = self.source._all_possible_ids()
2619
source_ids_set = set(source_ids)
2620
# source_ids is the worst possible case we may need to pull.
2621
# now we want to filter source_ids against what we actually
2622
# have in target, but don't try to check for existence where we know
2623
# we do not have a revision as that would be pointless.
2624
target_ids = set(self.target._all_possible_ids())
2625
possibly_present_revisions = target_ids.intersection(source_ids_set)
2626
actually_present_revisions = set(
2627
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2628
required_revisions = source_ids_set.difference(actually_present_revisions)
2629
if revision_id is not None:
2630
# we used get_ancestry to determine source_ids then we are assured all
2631
# revisions referenced are present as they are installed in topological order.
2632
# and the tip revision was validated by get_ancestry.
2633
result_set = required_revisions
2635
# if we just grabbed the possibly available ids, then
2636
# we only have an estimate of whats available and need to validate
2637
# that against the revision records.
2639
self.source._eliminate_revisions_not_present(required_revisions))
2640
return self.source.revision_ids_to_search_result(result_set)
2643
class InterKnitRepo(InterSameDataRepository):
2644
"""Optimised code paths between Knit based repositories."""
2647
def _get_repo_format_to_test(self):
2648
from bzrlib.repofmt import knitrepo
2649
return knitrepo.RepositoryFormatKnit1()
2652
def is_compatible(source, target):
2653
"""Be compatible with known Knit formats.
2655
We don't test for the stores being of specific types because that
2656
could lead to confusing results, and there is no need to be
2659
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2661
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2662
isinstance(target._format, RepositoryFormatKnit))
2663
except AttributeError:
2665
return are_knits and InterRepository._same_model(source, target)
2668
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2669
"""See InterRepository.fetch()."""
2670
from bzrlib.fetch import KnitRepoFetcher
2671
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2672
self.source, self.source._format, self.target, self.target._format)
2673
f = KnitRepoFetcher(to_repository=self.target,
2674
from_repository=self.source,
2675
last_revision=revision_id,
2676
pb=pb, find_ghosts=find_ghosts)
2677
return f.count_copied, f.failed_revisions
2680
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2681
"""See InterRepository.missing_revision_ids()."""
2682
if revision_id is not None:
2683
source_ids = self.source.get_ancestry(revision_id)
2684
if source_ids[0] is not None:
2685
raise AssertionError()
2688
source_ids = self.source.all_revision_ids()
2689
source_ids_set = set(source_ids)
2690
# source_ids is the worst possible case we may need to pull.
2691
# now we want to filter source_ids against what we actually
2692
# have in target, but don't try to check for existence where we know
2693
# we do not have a revision as that would be pointless.
2694
target_ids = set(self.target.all_revision_ids())
2695
possibly_present_revisions = target_ids.intersection(source_ids_set)
2696
actually_present_revisions = set(
2697
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2698
required_revisions = source_ids_set.difference(actually_present_revisions)
2699
if revision_id is not None:
2700
# we used get_ancestry to determine source_ids then we are assured all
2701
# revisions referenced are present as they are installed in topological order.
2702
# and the tip revision was validated by get_ancestry.
2703
result_set = required_revisions
2705
# if we just grabbed the possibly available ids, then
2706
# we only have an estimate of whats available and need to validate
2707
# that against the revision records.
2709
self.source._eliminate_revisions_not_present(required_revisions))
2710
return self.source.revision_ids_to_search_result(result_set)
2713
class InterPackRepo(InterSameDataRepository):
2714
"""Optimised code paths between Pack based repositories."""
2717
def _get_repo_format_to_test(self):
2718
from bzrlib.repofmt import pack_repo
2719
return pack_repo.RepositoryFormatKnitPack1()
2722
def is_compatible(source, target):
2723
"""Be compatible with known Pack formats.
2725
We don't test for the stores being of specific types because that
2726
could lead to confusing results, and there is no need to be
2729
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2731
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2732
isinstance(target._format, RepositoryFormatPack))
2733
except AttributeError:
2735
return are_packs and InterRepository._same_model(source, target)
2738
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2739
"""See InterRepository.fetch()."""
2740
from bzrlib.repofmt.pack_repo import Packer
2741
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2742
self.source, self.source._format, self.target, self.target._format)
2743
self.count_copied = 0
2744
if revision_id is None:
2746
# everything to do - use pack logic
2747
# to fetch from all packs to one without
2748
# inventory parsing etc, IFF nothing to be copied is in the target.
2750
revision_ids = self.source.all_revision_ids()
2751
# implementing the TODO will involve:
2752
# - detecting when all of a pack is selected
2753
# - avoiding as much as possible pre-selection, so the
2754
# more-core routines such as create_pack_from_packs can filter in
2755
# a just-in-time fashion. (though having a HEADS list on a
2756
# repository might make this a lot easier, because we could
2757
# sensibly detect 'new revisions' without doing a full index scan.
2758
elif _mod_revision.is_null(revision_id):
2763
revision_ids = self.search_missing_revision_ids(revision_id,
2764
find_ghosts=find_ghosts).get_keys()
2765
except errors.NoSuchRevision:
2766
raise errors.InstallFailed([revision_id])
2767
packs = self.source._pack_collection.all_packs()
2768
pack = Packer(self.target._pack_collection, packs, '.fetch',
2769
revision_ids).pack()
2770
if pack is not None:
2771
self.target._pack_collection._save_pack_names()
2772
# Trigger an autopack. This may duplicate effort as we've just done
2773
# a pack creation, but for now it is simpler to think about as
2774
# 'upload data, then repack if needed'.
2775
self.target._pack_collection.autopack()
2776
return (pack.get_revision_count(), [])
2781
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2782
"""See InterRepository.missing_revision_ids().
2784
:param find_ghosts: Find ghosts throughout the ancestry of
2787
if not find_ghosts and revision_id is not None:
2788
return self._walk_to_common_revisions([revision_id])
2789
elif revision_id is not None:
2790
source_ids = self.source.get_ancestry(revision_id)
2791
if source_ids[0] is not None:
2792
raise AssertionError()
2795
source_ids = self.source.all_revision_ids()
2796
# source_ids is the worst possible case we may need to pull.
2797
# now we want to filter source_ids against what we actually
2798
# have in target, but don't try to check for existence where we know
2799
# we do not have a revision as that would be pointless.
2800
target_ids = set(self.target.all_revision_ids())
2801
result_set = set(source_ids).difference(target_ids)
2802
return self.source.revision_ids_to_search_result(result_set)
2805
class InterModel1and2(InterRepository):
2808
def _get_repo_format_to_test(self):
2812
def is_compatible(source, target):
2813
if not source.supports_rich_root() and target.supports_rich_root():
2819
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2820
"""See InterRepository.fetch()."""
2821
from bzrlib.fetch import Model1toKnit2Fetcher
2822
f = Model1toKnit2Fetcher(to_repository=self.target,
2823
from_repository=self.source,
2824
last_revision=revision_id,
2825
pb=pb, find_ghosts=find_ghosts)
2826
return f.count_copied, f.failed_revisions
2829
def copy_content(self, revision_id=None):
2830
"""Make a complete copy of the content in self into destination.
2832
This is a destructive operation! Do not use it on existing
2835
:param revision_id: Only copy the content needed to construct
2836
revision_id and its parents.
2839
self.target.set_make_working_trees(self.source.make_working_trees())
2840
except NotImplementedError:
2842
# but don't bother fetching if we have the needed data now.
2843
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2844
self.target.has_revision(revision_id)):
2846
self.target.fetch(self.source, revision_id=revision_id)
2849
class InterKnit1and2(InterKnitRepo):
2852
def _get_repo_format_to_test(self):
2856
def is_compatible(source, target):
2857
"""Be compatible with Knit1 source and Knit3 target"""
2858
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2860
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2861
RepositoryFormatKnit3)
2862
from bzrlib.repofmt.pack_repo import (
2863
RepositoryFormatKnitPack1,
2864
RepositoryFormatKnitPack3,
2865
RepositoryFormatPackDevelopment0,
2866
RepositoryFormatPackDevelopment0Subtree,
2869
RepositoryFormatKnit1,
2870
RepositoryFormatKnitPack1,
2871
RepositoryFormatPackDevelopment0,
2874
RepositoryFormatKnit3,
2875
RepositoryFormatKnitPack3,
2876
RepositoryFormatPackDevelopment0Subtree,
2878
return (isinstance(source._format, nosubtrees) and
2879
isinstance(target._format, subtrees))
2880
except AttributeError:
2884
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2885
"""See InterRepository.fetch()."""
2886
from bzrlib.fetch import Knit1to2Fetcher
2887
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2888
self.source, self.source._format, self.target,
2889
self.target._format)
2890
f = Knit1to2Fetcher(to_repository=self.target,
2891
from_repository=self.source,
2892
last_revision=revision_id,
2893
pb=pb, find_ghosts=find_ghosts)
2894
return f.count_copied, f.failed_revisions
2897
class InterDifferingSerializer(InterKnitRepo):
2900
def _get_repo_format_to_test(self):
2904
def is_compatible(source, target):
2905
"""Be compatible with Knit2 source and Knit3 target"""
2906
if source.supports_rich_root() != target.supports_rich_root():
2908
# Ideally, we'd support fetching if the source had no tree references
2909
# even if it supported them...
2910
if (getattr(source, '_format.supports_tree_reference', False) and
2911
not getattr(target, '_format.supports_tree_reference', False)):
2916
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2917
"""See InterRepository.fetch()."""
2918
revision_ids = self.target.search_missing_revision_ids(self.source,
2919
revision_id, find_ghosts=find_ghosts).get_keys()
2920
revision_ids = tsort.topo_sort(
2921
self.source.get_graph().get_parent_map(revision_ids))
2922
def revisions_iterator():
2923
for current_revision_id in revision_ids:
2924
revision = self.source.get_revision(current_revision_id)
2925
tree = self.source.revision_tree(current_revision_id)
2927
signature = self.source.get_signature_text(
2928
current_revision_id)
2929
except errors.NoSuchRevision:
2931
yield revision, tree, signature
2933
my_pb = ui.ui_factory.nested_progress_bar()
2938
install_revisions(self.target, revisions_iterator(),
2939
len(revision_ids), pb)
2941
if my_pb is not None:
2943
return len(revision_ids), 0
2946
class InterRemoteToOther(InterRepository):
2948
def __init__(self, source, target):
2949
InterRepository.__init__(self, source, target)
2950
self._real_inter = None
2953
def is_compatible(source, target):
2954
if not isinstance(source, remote.RemoteRepository):
2956
# Is source's model compatible with target's model?
2957
source._ensure_real()
2958
real_source = source._real_repository
2959
if isinstance(real_source, remote.RemoteRepository):
2960
raise NotImplementedError(
2961
"We don't support remote repos backed by remote repos yet.")
2962
return InterRepository._same_model(real_source, target)
2965
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2966
"""See InterRepository.fetch()."""
2967
from bzrlib.fetch import RemoteToOtherFetcher
2968
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2969
self.source, self.target, self.target._format)
2970
# TODO: jam 20070210 This should be an assert, not a translate
2971
revision_id = osutils.safe_revision_id(revision_id)
2972
f = RemoteToOtherFetcher(to_repository=self.target,
2973
from_repository=self.source,
2974
last_revision=revision_id,
2975
pb=pb, find_ghosts=find_ghosts)
2976
return f.count_copied, f.failed_revisions
2979
def _get_repo_format_to_test(self):
2983
class InterOtherToRemote(InterRepository):
2985
def __init__(self, source, target):
2986
InterRepository.__init__(self, source, target)
2987
self._real_inter = None
2990
def is_compatible(source, target):
2991
if isinstance(target, remote.RemoteRepository):
2995
def _ensure_real_inter(self):
2996
if self._real_inter is None:
2997
self.target._ensure_real()
2998
real_target = self.target._real_repository
2999
self._real_inter = InterRepository.get(self.source, real_target)
3001
def copy_content(self, revision_id=None):
3002
self._ensure_real_inter()
3003
self._real_inter.copy_content(revision_id=revision_id)
3005
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
3006
self._ensure_real_inter()
3007
self._real_inter.fetch(revision_id=revision_id, pb=pb,
3008
find_ghosts=find_ghosts)
3011
def _get_repo_format_to_test(self):
3015
InterRepository.register_optimiser(InterDifferingSerializer)
3016
InterRepository.register_optimiser(InterSameDataRepository)
3017
InterRepository.register_optimiser(InterWeaveRepo)
3018
InterRepository.register_optimiser(InterKnitRepo)
3019
InterRepository.register_optimiser(InterModel1and2)
3020
InterRepository.register_optimiser(InterKnit1and2)
3021
InterRepository.register_optimiser(InterPackRepo)
3022
InterRepository.register_optimiser(InterRemoteToOther)
3023
InterRepository.register_optimiser(InterOtherToRemote)
3026
class CopyConverter(object):
3027
"""A repository conversion tool which just performs a copy of the content.
3029
This is slow but quite reliable.
3032
def __init__(self, target_format):
3033
"""Create a CopyConverter.
3035
:param target_format: The format the resulting repository should be.
3037
self.target_format = target_format
3039
def convert(self, repo, pb):
3040
"""Perform the conversion of to_convert, giving feedback via pb.
3042
:param to_convert: The disk object to convert.
3043
:param pb: a progress bar to use for progress information.
3048
# this is only useful with metadir layouts - separated repo content.
3049
# trigger an assertion if not such
3050
repo._format.get_format_string()
3051
self.repo_dir = repo.bzrdir
3052
self.step('Moving repository to repository.backup')
3053
self.repo_dir.transport.move('repository', 'repository.backup')
3054
backup_transport = self.repo_dir.transport.clone('repository.backup')
3055
repo._format.check_conversion_target(self.target_format)
3056
self.source_repo = repo._format.open(self.repo_dir,
3058
_override_transport=backup_transport)
3059
self.step('Creating new repository')
3060
converted = self.target_format.initialize(self.repo_dir,
3061
self.source_repo.is_shared())
3062
converted.lock_write()
3064
self.step('Copying content into repository.')
3065
self.source_repo.copy_content_into(converted)
3068
self.step('Deleting old repository content.')
3069
self.repo_dir.transport.delete_tree('repository.backup')
3070
self.pb.note('repository converted')
3072
def step(self, message):
3073
"""Update the pb by a step."""
3075
self.pb.update(message, self.count, self.total)
3087
def _unescaper(match, _map=_unescape_map):
3088
code = match.group(1)
3092
if not code.startswith('#'):
3094
return unichr(int(code[1:])).encode('utf8')
3100
def _unescape_xml(data):
3101
"""Unescape predefined XML entities in a string of data."""
3103
if _unescape_re is None:
3104
_unescape_re = re.compile('\&([^;]*);')
3105
return _unescape_re.sub(_unescaper, data)
3108
class _VersionedFileChecker(object):
3110
def __init__(self, repository):
3111
self.repository = repository
3112
self.text_index = self.repository._generate_text_key_index()
3114
def calculate_file_version_parents(self, revision_id, file_id):
3115
"""Calculate the correct parents for a file version according to
3118
parent_keys = self.text_index[(file_id, revision_id)]
3119
if parent_keys == [_mod_revision.NULL_REVISION]:
3121
# strip the file_id, for the weave api
3122
return tuple([revision_id for file_id, revision_id in parent_keys])
3124
def check_file_version_parents(self, weave, file_id):
3125
"""Check the parents stored in a versioned file are correct.
3127
It also detects file versions that are not referenced by their
3128
corresponding revision's inventory.
3130
:returns: A tuple of (wrong_parents, dangling_file_versions).
3131
wrong_parents is a dict mapping {revision_id: (stored_parents,
3132
correct_parents)} for each revision_id where the stored parents
3133
are not correct. dangling_file_versions is a set of (file_id,
3134
revision_id) tuples for versions that are present in this versioned
3135
file, but not used by the corresponding inventory.
3138
unused_versions = set()
3139
versions = weave.versions()
3140
parent_map = weave.get_parent_map(versions)
3141
for num, revision_id in enumerate(versions):
3143
correct_parents = self.calculate_file_version_parents(
3144
revision_id, file_id)
3146
# The version is not part of the used keys.
3147
unused_versions.add(revision_id)
3150
knit_parents = tuple(parent_map[revision_id])
3151
except errors.RevisionNotPresent:
3153
if correct_parents != knit_parents:
3154
wrong_parents[revision_id] = (knit_parents, correct_parents)
3155
return wrong_parents, unused_versions
3158
def _old_get_graph(repository, revision_id):
3159
"""DO NOT USE. That is all. I'm serious."""
3160
graph = repository.get_graph()
3161
revision_graph = dict(((key, value) for key, value in
3162
graph.iter_ancestry([revision_id]) if value is not None))
3163
return _strip_NULL_ghosts(revision_graph)
3166
def _strip_NULL_ghosts(revision_graph):
3167
"""Also don't use this. more compatibility code for unmigrated clients."""
3168
# Filter ghosts, and null:
3169
if _mod_revision.NULL_REVISION in revision_graph:
3170
del revision_graph[_mod_revision.NULL_REVISION]
3171
for key, parents in revision_graph.items():
3172
revision_graph[key] = tuple(parent for parent in parents if parent
3174
return revision_graph
added
removed
bzrlib/repository.py
