1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
39
revision as _mod_revision,
45
from bzrlib.bundle import serializer
46
from bzrlib.revisiontree import RevisionTree
47
from bzrlib.store.versioned import VersionedFileStore
48
from bzrlib.store.text import TextStore
49
from bzrlib.testament import Testament
50
from bzrlib.util import bencode
53
from bzrlib.decorators import needs_read_lock, needs_write_lock
54
from bzrlib.inter import InterObject
55
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
56
from bzrlib.symbol_versioning import (
59
from bzrlib.trace import mutter, mutter_callsite, note, warning
62
# Old formats display a warning, but only once
63
_deprecation_warning_done = False
66
class CommitBuilder(object):
67
"""Provides an interface to build up a commit.
69
This allows describing a tree to be committed without needing to
70
know the internals of the format of the repository.
73
# all clients should supply tree roots.
74
record_root_entry = True
75
# the default CommitBuilder does not manage trees whose root is versioned.
76
_versioned_root = False
78
def __init__(self, repository, parents, config, timestamp=None,
79
timezone=None, committer=None, revprops=None,
81
"""Initiate a CommitBuilder.
83
:param repository: Repository to commit to.
84
:param parents: Revision ids of the parents of the new revision.
85
:param config: Configuration to use.
86
:param timestamp: Optional timestamp recorded for commit.
87
:param timezone: Optional timezone for timestamp.
88
:param committer: Optional committer to set for commit.
89
:param revprops: Optional dictionary of revision properties.
90
:param revision_id: Optional revision id.
95
self._committer = self._config.username()
97
self._committer = committer
99
self.new_inventory = Inventory(None)
100
self._new_revision_id = revision_id
101
self.parents = parents
102
self.repository = repository
105
if revprops is not None:
106
self._revprops.update(revprops)
108
if timestamp is None:
109
timestamp = time.time()
110
# Restrict resolution to 1ms
111
self._timestamp = round(timestamp, 3)
114
self._timezone = osutils.local_time_offset()
116
self._timezone = int(timezone)
118
self._generate_revision_if_needed()
119
self.__heads = graph.HeadsCache(repository.get_graph()).heads
121
def commit(self, message):
122
"""Make the actual commit.
124
:return: The revision id of the recorded revision.
126
rev = _mod_revision.Revision(
127
timestamp=self._timestamp,
128
timezone=self._timezone,
129
committer=self._committer,
131
inventory_sha1=self.inv_sha1,
132
revision_id=self._new_revision_id,
133
properties=self._revprops)
134
rev.parent_ids = self.parents
135
self.repository.add_revision(self._new_revision_id, rev,
136
self.new_inventory, self._config)
137
self.repository.commit_write_group()
138
return self._new_revision_id
141
"""Abort the commit that is being built.
143
self.repository.abort_write_group()
145
def revision_tree(self):
146
"""Return the tree that was just committed.
148
After calling commit() this can be called to get a RevisionTree
149
representing the newly committed tree. This is preferred to
150
calling Repository.revision_tree() because that may require
151
deserializing the inventory, while we already have a copy in
154
return RevisionTree(self.repository, self.new_inventory,
155
self._new_revision_id)
157
def finish_inventory(self):
158
"""Tell the builder that the inventory is finished."""
159
if self.new_inventory.root is None:
160
raise AssertionError('Root entry should be supplied to'
161
' record_entry_contents, as of bzr 0.10.',
162
DeprecationWarning, stacklevel=2)
163
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
164
self.new_inventory.revision_id = self._new_revision_id
165
self.inv_sha1 = self.repository.add_inventory(
166
self._new_revision_id,
171
def _gen_revision_id(self):
172
"""Return new revision-id."""
173
return generate_ids.gen_revision_id(self._config.username(),
176
def _generate_revision_if_needed(self):
177
"""Create a revision id if None was supplied.
179
If the repository can not support user-specified revision ids
180
they should override this function and raise CannotSetRevisionId
181
if _new_revision_id is not None.
183
:raises: CannotSetRevisionId
185
if self._new_revision_id is None:
186
self._new_revision_id = self._gen_revision_id()
187
self.random_revid = True
189
self.random_revid = False
191
def _heads(self, file_id, revision_ids):
192
"""Calculate the graph heads for revision_ids in the graph of file_id.
194
This can use either a per-file graph or a global revision graph as we
195
have an identity relationship between the two graphs.
197
return self.__heads(revision_ids)
199
def _check_root(self, ie, parent_invs, tree):
200
"""Helper for record_entry_contents.
202
:param ie: An entry being added.
203
:param parent_invs: The inventories of the parent revisions of the
205
:param tree: The tree that is being committed.
207
# In this revision format, root entries have no knit or weave When
208
# serializing out to disk and back in root.revision is always
210
ie.revision = self._new_revision_id
212
def _get_delta(self, ie, basis_inv, path):
213
"""Get a delta against the basis inventory for ie."""
214
if ie.file_id not in basis_inv:
216
return (None, path, ie.file_id, ie)
217
elif ie != basis_inv[ie.file_id]:
219
# TODO: avoid tis id2path call.
220
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
225
def record_entry_contents(self, ie, parent_invs, path, tree,
227
"""Record the content of ie from tree into the commit if needed.
229
Side effect: sets ie.revision when unchanged
231
:param ie: An inventory entry present in the commit.
232
:param parent_invs: The inventories of the parent revisions of the
234
:param path: The path the entry is at in the tree.
235
:param tree: The tree which contains this entry and should be used to
237
:param content_summary: Summary data from the tree about the paths
238
content - stat, length, exec, sha/link target. This is only
239
accessed when the entry has a revision of None - that is when it is
240
a candidate to commit.
241
:return: A tuple (change_delta, version_recorded). change_delta is
242
an inventory_delta change for this entry against the basis tree of
243
the commit, or None if no change occured against the basis tree.
244
version_recorded is True if a new version of the entry has been
245
recorded. For instance, committing a merge where a file was only
246
changed on the other side will return (delta, False).
248
if self.new_inventory.root is None:
249
if ie.parent_id is not None:
250
raise errors.RootMissing()
251
self._check_root(ie, parent_invs, tree)
252
if ie.revision is None:
253
kind = content_summary[0]
255
# ie is carried over from a prior commit
257
# XXX: repository specific check for nested tree support goes here - if
258
# the repo doesn't want nested trees we skip it ?
259
if (kind == 'tree-reference' and
260
not self.repository._format.supports_tree_reference):
261
# mismatch between commit builder logic and repository:
262
# this needs the entry creation pushed down into the builder.
263
raise NotImplementedError('Missing repository subtree support.')
264
self.new_inventory.add(ie)
266
# TODO: slow, take it out of the inner loop.
268
basis_inv = parent_invs[0]
270
basis_inv = Inventory(root_id=None)
272
# ie.revision is always None if the InventoryEntry is considered
273
# for committing. We may record the previous parents revision if the
274
# content is actually unchanged against a sole head.
275
if ie.revision is not None:
276
if not self._versioned_root and path == '':
277
# repositories that do not version the root set the root's
278
# revision to the new commit even when no change occurs, and
279
# this masks when a change may have occurred against the basis,
280
# so calculate if one happened.
281
if ie.file_id in basis_inv:
282
delta = (basis_inv.id2path(ie.file_id), path,
286
delta = (None, path, ie.file_id, ie)
289
# we don't need to commit this, because the caller already
290
# determined that an existing revision of this file is
292
return None, (ie.revision == self._new_revision_id)
293
# XXX: Friction: parent_candidates should return a list not a dict
294
# so that we don't have to walk the inventories again.
295
parent_candiate_entries = ie.parent_candidates(parent_invs)
296
head_set = self._heads(ie.file_id, parent_candiate_entries.keys())
298
for inv in parent_invs:
299
if ie.file_id in inv:
300
old_rev = inv[ie.file_id].revision
301
if old_rev in head_set:
302
heads.append(inv[ie.file_id].revision)
303
head_set.remove(inv[ie.file_id].revision)
306
# now we check to see if we need to write a new record to the
308
# We write a new entry unless there is one head to the ancestors, and
309
# the kind-derived content is unchanged.
311
# Cheapest check first: no ancestors, or more the one head in the
312
# ancestors, we write a new node.
316
# There is a single head, look it up for comparison
317
parent_entry = parent_candiate_entries[heads[0]]
318
# if the non-content specific data has changed, we'll be writing a
320
if (parent_entry.parent_id != ie.parent_id or
321
parent_entry.name != ie.name):
323
# now we need to do content specific checks:
325
# if the kind changed the content obviously has
326
if kind != parent_entry.kind:
329
if content_summary[2] is None:
330
raise ValueError("Files must not have executable = None")
332
if (# if the file length changed we have to store:
333
parent_entry.text_size != content_summary[1] or
334
# if the exec bit has changed we have to store:
335
parent_entry.executable != content_summary[2]):
337
elif parent_entry.text_sha1 == content_summary[3]:
338
# all meta and content is unchanged (using a hash cache
339
# hit to check the sha)
340
ie.revision = parent_entry.revision
341
ie.text_size = parent_entry.text_size
342
ie.text_sha1 = parent_entry.text_sha1
343
ie.executable = parent_entry.executable
344
return self._get_delta(ie, basis_inv, path), False
346
# Either there is only a hash change(no hash cache entry,
347
# or same size content change), or there is no change on
349
# Provide the parent's hash to the store layer, so that the
350
# content is unchanged we will not store a new node.
351
nostore_sha = parent_entry.text_sha1
353
# We want to record a new node regardless of the presence or
354
# absence of a content change in the file.
356
ie.executable = content_summary[2]
357
lines = tree.get_file(ie.file_id, path).readlines()
359
ie.text_sha1, ie.text_size = self._add_text_to_weave(
360
ie.file_id, lines, heads, nostore_sha)
361
except errors.ExistingContent:
362
# Turns out that the file content was unchanged, and we were
363
# only going to store a new node if it was changed. Carry over
365
ie.revision = parent_entry.revision
366
ie.text_size = parent_entry.text_size
367
ie.text_sha1 = parent_entry.text_sha1
368
ie.executable = parent_entry.executable
369
return self._get_delta(ie, basis_inv, path), False
370
elif kind == 'directory':
372
# all data is meta here, nothing specific to directory, so
374
ie.revision = parent_entry.revision
375
return self._get_delta(ie, basis_inv, path), False
377
self._add_text_to_weave(ie.file_id, lines, heads, None)
378
elif kind == 'symlink':
379
current_link_target = content_summary[3]
381
# symlink target is not generic metadata, check if it has
383
if current_link_target != parent_entry.symlink_target:
386
# unchanged, carry over.
387
ie.revision = parent_entry.revision
388
ie.symlink_target = parent_entry.symlink_target
389
return self._get_delta(ie, basis_inv, path), False
390
ie.symlink_target = current_link_target
392
self._add_text_to_weave(ie.file_id, lines, heads, None)
393
elif kind == 'tree-reference':
395
if content_summary[3] != parent_entry.reference_revision:
398
# unchanged, carry over.
399
ie.reference_revision = parent_entry.reference_revision
400
ie.revision = parent_entry.revision
401
return self._get_delta(ie, basis_inv, path), False
402
ie.reference_revision = content_summary[3]
404
self._add_text_to_weave(ie.file_id, lines, heads, None)
406
raise NotImplementedError('unknown kind')
407
ie.revision = self._new_revision_id
408
return self._get_delta(ie, basis_inv, path), True
410
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
411
versionedfile = self.repository.weave_store.get_weave_or_empty(
412
file_id, self.repository.get_transaction())
413
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
414
# than add_lines, and allows committing when a parent is ghosted for
416
# Note: as we read the content directly from the tree, we know its not
417
# been turned into unicode or badly split - but a broken tree
418
# implementation could give us bad output from readlines() so this is
419
# not a guarantee of safety. What would be better is always checking
420
# the content during test suite execution. RBC 20070912
421
return versionedfile.add_lines_with_ghosts(
422
self._new_revision_id, parents, new_lines,
423
nostore_sha=nostore_sha, random_id=self.random_revid,
424
check_content=False)[0:2]
427
class RootCommitBuilder(CommitBuilder):
428
"""This commitbuilder actually records the root id"""
430
# the root entry gets versioned properly by this builder.
431
_versioned_root = True
433
def _check_root(self, ie, parent_invs, tree):
434
"""Helper for record_entry_contents.
436
:param ie: An entry being added.
437
:param parent_invs: The inventories of the parent revisions of the
439
:param tree: The tree that is being committed.
443
######################################################################
446
class Repository(object):
447
"""Repository holding history for one or more branches.
449
The repository holds and retrieves historical information including
450
revisions and file history. It's normally accessed only by the Branch,
451
which views a particular line of development through that history.
453
The Repository builds on top of Stores and a Transport, which respectively
454
describe the disk data format and the way of accessing the (possibly
458
# What class to use for a CommitBuilder. Often its simpler to change this
459
# in a Repository class subclass rather than to override
460
# get_commit_builder.
461
_commit_builder_class = CommitBuilder
462
# The search regex used by xml based repositories to determine what things
463
# where changed in a single commit.
464
_file_ids_altered_regex = lazy_regex.lazy_compile(
465
r'file_id="(?P<file_id>[^"]+)"'
466
r'.* revision="(?P<revision_id>[^"]+)"'
469
def abort_write_group(self):
470
"""Commit the contents accrued within the current write group.
472
:seealso: start_write_group.
474
if self._write_group is not self.get_transaction():
475
# has an unlock or relock occured ?
476
raise errors.BzrError('mismatched lock context and write group.')
477
self._abort_write_group()
478
self._write_group = None
480
def _abort_write_group(self):
481
"""Template method for per-repository write group cleanup.
483
This is called during abort before the write group is considered to be
484
finished and should cleanup any internal state accrued during the write
485
group. There is no requirement that data handed to the repository be
486
*not* made available - this is not a rollback - but neither should any
487
attempt be made to ensure that data added is fully commited. Abort is
488
invoked when an error has occured so futher disk or network operations
489
may not be possible or may error and if possible should not be
493
def add_inventory(self, revision_id, inv, parents):
494
"""Add the inventory inv to the repository as revision_id.
496
:param parents: The revision ids of the parents that revision_id
497
is known to have and are in the repository already.
499
:returns: The validator(which is a sha1 digest, though what is sha'd is
500
repository format specific) of the serialized inventory.
502
if not self.is_in_write_group():
503
raise AssertionError("%r not in write group" % (self,))
504
_mod_revision.check_not_reserved_id(revision_id)
505
if not (inv.revision_id is None or inv.revision_id == revision_id):
506
raise AssertionError(
507
"Mismatch between inventory revision"
508
" id and insertion revid (%r, %r)"
509
% (inv.revision_id, revision_id))
511
raise AssertionError()
512
inv_lines = self._serialise_inventory_to_lines(inv)
513
inv_vf = self.get_inventory_weave()
514
return self._inventory_add_lines(inv_vf, revision_id, parents,
515
inv_lines, check_content=False)
517
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
519
"""Store lines in inv_vf and return the sha1 of the inventory."""
521
for parent in parents:
523
final_parents.append(parent)
524
return inv_vf.add_lines(revision_id, final_parents, lines,
525
check_content=check_content)[0]
527
def add_revision(self, revision_id, rev, inv=None, config=None):
528
"""Add rev to the revision store as revision_id.
530
:param revision_id: the revision id to use.
531
:param rev: The revision object.
532
:param inv: The inventory for the revision. if None, it will be looked
533
up in the inventory storer
534
:param config: If None no digital signature will be created.
535
If supplied its signature_needed method will be used
536
to determine if a signature should be made.
538
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
540
_mod_revision.check_not_reserved_id(revision_id)
541
if config is not None and config.signature_needed():
543
inv = self.get_inventory(revision_id)
544
plaintext = Testament(rev, inv).as_short_text()
545
self.store_revision_signature(
546
gpg.GPGStrategy(config), plaintext, revision_id)
547
inventory_vf = self.get_inventory_weave()
548
if not revision_id in inventory_vf:
550
raise errors.WeaveRevisionNotPresent(revision_id,
553
# yes, this is not suitable for adding with ghosts.
554
rev.inventory_sha1 = self.add_inventory(revision_id, inv,
557
rev.inventory_sha1 = inventory_vf.get_sha1s([revision_id])[0]
558
self._revision_store.add_revision(rev, self.get_transaction())
560
def _add_revision_text(self, revision_id, text):
561
revision = self._revision_store._serializer.read_revision_from_string(
563
self._revision_store._add_revision(revision, StringIO(text),
564
self.get_transaction())
566
def all_revision_ids(self):
567
"""Returns a list of all the revision ids in the repository.
569
This is deprecated because code should generally work on the graph
570
reachable from a particular revision, and ignore any other revisions
571
that might be present. There is no direct replacement method.
573
if 'evil' in debug.debug_flags:
574
mutter_callsite(2, "all_revision_ids is linear with history.")
575
return self._all_revision_ids()
577
def _all_revision_ids(self):
578
"""Returns a list of all the revision ids in the repository.
580
These are in as much topological order as the underlying store can
583
raise NotImplementedError(self._all_revision_ids)
585
def break_lock(self):
586
"""Break a lock if one is present from another instance.
588
Uses the ui factory to ask for confirmation if the lock may be from
591
self.control_files.break_lock()
594
def _eliminate_revisions_not_present(self, revision_ids):
595
"""Check every revision id in revision_ids to see if we have it.
597
Returns a set of the present revisions.
600
graph = self.get_graph()
601
parent_map = graph.get_parent_map(revision_ids)
602
# The old API returned a list, should this actually be a set?
603
return parent_map.keys()
606
def create(a_bzrdir):
607
"""Construct the current default format repository in a_bzrdir."""
608
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
610
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
611
"""instantiate a Repository.
613
:param _format: The format of the repository on disk.
614
:param a_bzrdir: The BzrDir of the repository.
616
In the future we will have a single api for all stores for
617
getting file texts, inventories and revisions, then
618
this construct will accept instances of those things.
620
super(Repository, self).__init__()
621
self._format = _format
622
# the following are part of the public API for Repository:
623
self.bzrdir = a_bzrdir
624
self.control_files = control_files
625
self._revision_store = _revision_store
626
# backwards compatibility
627
self.weave_store = text_store
629
self._reconcile_does_inventory_gc = True
630
self._reconcile_fixes_text_parents = False
631
self._reconcile_backsup_inventory = True
632
# not right yet - should be more semantically clear ?
634
self.control_store = control_store
635
self.control_weaves = control_store
636
# TODO: make sure to construct the right store classes, etc, depending
637
# on whether escaping is required.
638
self._warn_if_deprecated()
639
self._write_group = None
640
self.base = control_files._transport.base
643
return '%s(%r)' % (self.__class__.__name__,
646
def has_same_location(self, other):
647
"""Returns a boolean indicating if this repository is at the same
648
location as another repository.
650
This might return False even when two repository objects are accessing
651
the same physical repository via different URLs.
653
if self.__class__ is not other.__class__:
655
return (self.control_files._transport.base ==
656
other.control_files._transport.base)
658
def is_in_write_group(self):
659
"""Return True if there is an open write group.
661
:seealso: start_write_group.
663
return self._write_group is not None
666
return self.control_files.is_locked()
668
def is_write_locked(self):
669
"""Return True if this object is write locked."""
670
return self.is_locked() and self.control_files._lock_mode == 'w'
672
def lock_write(self, token=None):
673
"""Lock this repository for writing.
675
This causes caching within the repository obejct to start accumlating
676
data during reads, and allows a 'write_group' to be obtained. Write
677
groups must be used for actual data insertion.
679
:param token: if this is already locked, then lock_write will fail
680
unless the token matches the existing lock.
681
:returns: a token if this instance supports tokens, otherwise None.
682
:raises TokenLockingNotSupported: when a token is given but this
683
instance doesn't support using token locks.
684
:raises MismatchedToken: if the specified token doesn't match the token
685
of the existing lock.
686
:seealso: start_write_group.
688
A token should be passed in if you know that you have locked the object
689
some other way, and need to synchronise this object's state with that
692
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
694
result = self.control_files.lock_write(token=token)
699
self.control_files.lock_read()
702
def get_physical_lock_status(self):
703
return self.control_files.get_physical_lock_status()
705
def leave_lock_in_place(self):
706
"""Tell this repository not to release the physical lock when this
709
If lock_write doesn't return a token, then this method is not supported.
711
self.control_files.leave_in_place()
713
def dont_leave_lock_in_place(self):
714
"""Tell this repository to release the physical lock when this
715
object is unlocked, even if it didn't originally acquire it.
717
If lock_write doesn't return a token, then this method is not supported.
719
self.control_files.dont_leave_in_place()
722
def gather_stats(self, revid=None, committers=None):
723
"""Gather statistics from a revision id.
725
:param revid: The revision id to gather statistics from, if None, then
726
no revision specific statistics are gathered.
727
:param committers: Optional parameter controlling whether to grab
728
a count of committers from the revision specific statistics.
729
:return: A dictionary of statistics. Currently this contains:
730
committers: The number of committers if requested.
731
firstrev: A tuple with timestamp, timezone for the penultimate left
732
most ancestor of revid, if revid is not the NULL_REVISION.
733
latestrev: A tuple with timestamp, timezone for revid, if revid is
734
not the NULL_REVISION.
735
revisions: The total revision count in the repository.
736
size: An estimate disk size of the repository in bytes.
739
if revid and committers:
740
result['committers'] = 0
741
if revid and revid != _mod_revision.NULL_REVISION:
743
all_committers = set()
744
revisions = self.get_ancestry(revid)
745
# pop the leading None
747
first_revision = None
749
# ignore the revisions in the middle - just grab first and last
750
revisions = revisions[0], revisions[-1]
751
for revision in self.get_revisions(revisions):
752
if not first_revision:
753
first_revision = revision
755
all_committers.add(revision.committer)
756
last_revision = revision
758
result['committers'] = len(all_committers)
759
result['firstrev'] = (first_revision.timestamp,
760
first_revision.timezone)
761
result['latestrev'] = (last_revision.timestamp,
762
last_revision.timezone)
764
# now gather global repository information
765
if self.bzrdir.root_transport.listable():
766
c, t = self._revision_store.total_size(self.get_transaction())
767
result['revisions'] = c
771
def find_branches(self, using=False):
772
"""Find branches underneath this repository.
774
This will include branches inside other branches.
776
:param using: If True, list only branches using this repository.
778
if using and not self.is_shared():
780
return [self.bzrdir.open_branch()]
781
except errors.NotBranchError:
783
class Evaluator(object):
786
self.first_call = True
788
def __call__(self, bzrdir):
789
# On the first call, the parameter is always the bzrdir
790
# containing the current repo.
791
if not self.first_call:
793
repository = bzrdir.open_repository()
794
except errors.NoRepositoryPresent:
797
return False, (None, repository)
798
self.first_call = False
800
value = (bzrdir.open_branch(), None)
801
except errors.NotBranchError:
806
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
807
self.bzrdir.root_transport, evaluate=Evaluator()):
808
if branch is not None:
809
branches.append(branch)
810
if not using and repository is not None:
811
branches.extend(repository.find_branches())
814
def get_data_stream(self, revision_ids):
815
raise NotImplementedError(self.get_data_stream)
817
def get_data_stream_for_search(self, search_result):
818
"""Get a data stream that can be inserted to a repository.
820
:param search_result: A bzrlib.graph.SearchResult selecting the
822
:return: A data stream that can be inserted into a repository using
825
raise NotImplementedError(self.get_data_stream_for_search)
827
def insert_data_stream(self, stream):
828
"""XXX What does this really do?
830
Is it a substitute for fetch?
831
Should it manage its own write group ?
833
for item_key, bytes in stream:
834
if item_key[0] == 'file':
835
(file_id,) = item_key[1:]
836
knit = self.weave_store.get_weave_or_empty(
837
file_id, self.get_transaction())
838
elif item_key == ('inventory',):
839
knit = self.get_inventory_weave()
840
elif item_key == ('revisions',):
841
knit = self._revision_store.get_revision_file(
842
self.get_transaction())
843
elif item_key == ('signatures',):
844
knit = self._revision_store.get_signature_file(
845
self.get_transaction())
847
raise errors.RepositoryDataStreamError(
848
"Unrecognised data stream key '%s'" % (item_key,))
849
decoded_list = bencode.bdecode(bytes)
850
format = decoded_list.pop(0)
853
for version, options, parents, some_bytes in decoded_list:
854
data_list.append((version, options, len(some_bytes), parents))
855
knit_bytes += some_bytes
856
buffer = StringIO(knit_bytes)
857
def reader_func(count):
861
return buffer.read(count)
862
knit.insert_data_stream(
863
(format, data_list, reader_func))
866
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
867
"""Return the revision ids that other has that this does not.
869
These are returned in topological order.
871
revision_id: only return revision ids included by revision_id.
873
return InterRepository.get(other, self).search_missing_revision_ids(
874
revision_id, find_ghosts)
876
@deprecated_method(symbol_versioning.one_two)
878
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
879
"""Return the revision ids that other has that this does not.
881
These are returned in topological order.
883
revision_id: only return revision ids included by revision_id.
885
keys = self.search_missing_revision_ids(
886
other, revision_id, find_ghosts).get_keys()
889
parents = other.get_graph().get_parent_map(keys)
892
return tsort.topo_sort(parents)
896
"""Open the repository rooted at base.
898
For instance, if the repository is at URL/.bzr/repository,
899
Repository.open(URL) -> a Repository instance.
901
control = bzrdir.BzrDir.open(base)
902
return control.open_repository()
904
def copy_content_into(self, destination, revision_id=None):
905
"""Make a complete copy of the content in self into destination.
907
This is a destructive operation! Do not use it on existing
910
return InterRepository.get(self, destination).copy_content(revision_id)
912
def commit_write_group(self):
913
"""Commit the contents accrued within the current write group.
915
:seealso: start_write_group.
917
if self._write_group is not self.get_transaction():
918
# has an unlock or relock occured ?
919
raise errors.BzrError('mismatched lock context %r and '
921
(self.get_transaction(), self._write_group))
922
self._commit_write_group()
923
self._write_group = None
925
def _commit_write_group(self):
926
"""Template method for per-repository write group cleanup.
928
This is called before the write group is considered to be
929
finished and should ensure that all data handed to the repository
930
for writing during the write group is safely committed (to the
931
extent possible considering file system caching etc).
934
def fetch(self, source, revision_id=None, pb=None, find_ghosts=False):
935
"""Fetch the content required to construct revision_id from source.
937
If revision_id is None all content is copied.
938
:param find_ghosts: Find and copy revisions in the source that are
939
ghosts in the target (and not reachable directly by walking out to
940
the first-present revision in target from revision_id).
942
# fast path same-url fetch operations
943
if self.has_same_location(source):
944
# check that last_revision is in 'from' and then return a
946
if (revision_id is not None and
947
not _mod_revision.is_null(revision_id)):
948
self.get_revision(revision_id)
950
inter = InterRepository.get(source, self)
952
return inter.fetch(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts)
953
except NotImplementedError:
954
raise errors.IncompatibleRepositories(source, self)
956
def create_bundle(self, target, base, fileobj, format=None):
957
return serializer.write_bundle(self, target, base, fileobj, format)
959
def get_commit_builder(self, branch, parents, config, timestamp=None,
960
timezone=None, committer=None, revprops=None,
962
"""Obtain a CommitBuilder for this repository.
964
:param branch: Branch to commit to.
965
:param parents: Revision ids of the parents of the new revision.
966
:param config: Configuration to use.
967
:param timestamp: Optional timestamp recorded for commit.
968
:param timezone: Optional timezone for timestamp.
969
:param committer: Optional committer to set for commit.
970
:param revprops: Optional dictionary of revision properties.
971
:param revision_id: Optional revision id.
973
result = self._commit_builder_class(self, parents, config,
974
timestamp, timezone, committer, revprops, revision_id)
975
self.start_write_group()
979
if (self.control_files._lock_count == 1 and
980
self.control_files._lock_mode == 'w'):
981
if self._write_group is not None:
982
self.abort_write_group()
983
self.control_files.unlock()
984
raise errors.BzrError(
985
'Must end write groups before releasing write locks.')
986
self.control_files.unlock()
989
def clone(self, a_bzrdir, revision_id=None):
990
"""Clone this repository into a_bzrdir using the current format.
992
Currently no check is made that the format of this repository and
993
the bzrdir format are compatible. FIXME RBC 20060201.
995
:return: The newly created destination repository.
997
# TODO: deprecate after 0.16; cloning this with all its settings is
998
# probably not very useful -- mbp 20070423
999
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
1000
self.copy_content_into(dest_repo, revision_id)
1003
def start_write_group(self):
1004
"""Start a write group in the repository.
1006
Write groups are used by repositories which do not have a 1:1 mapping
1007
between file ids and backend store to manage the insertion of data from
1008
both fetch and commit operations.
1010
A write lock is required around the start_write_group/commit_write_group
1011
for the support of lock-requiring repository formats.
1013
One can only insert data into a repository inside a write group.
1017
if not self.is_write_locked():
1018
raise errors.NotWriteLocked(self)
1019
if self._write_group:
1020
raise errors.BzrError('already in a write group')
1021
self._start_write_group()
1022
# so we can detect unlock/relock - the write group is now entered.
1023
self._write_group = self.get_transaction()
1025
def _start_write_group(self):
1026
"""Template method for per-repository write group startup.
1028
This is called before the write group is considered to be
1033
def sprout(self, to_bzrdir, revision_id=None):
1034
"""Create a descendent repository for new development.
1036
Unlike clone, this does not copy the settings of the repository.
1038
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
1039
dest_repo.fetch(self, revision_id=revision_id)
1042
def _create_sprouting_repo(self, a_bzrdir, shared):
1043
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
1044
# use target default format.
1045
dest_repo = a_bzrdir.create_repository()
1047
# Most control formats need the repository to be specifically
1048
# created, but on some old all-in-one formats it's not needed
1050
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
1051
except errors.UninitializableFormat:
1052
dest_repo = a_bzrdir.open_repository()
1056
def has_revision(self, revision_id):
1057
"""True if this repository has a copy of the revision."""
1058
return revision_id in self.has_revisions((revision_id,))
1060
def has_revisions(self, revision_ids):
1061
"""Probe to find out the presence of multiple revisions.
1063
:param revision_ids: An iterable of revision_ids.
1064
:return: A set of the revision_ids that were present.
1066
raise NotImplementedError(self.has_revisions)
1068
return self._revision_store.has_revision_id(revision_id,
1069
self.get_transaction())
1072
def get_revision(self, revision_id):
1073
"""Return the Revision object for a named revision."""
1074
return self.get_revisions([revision_id])[0]
1077
def get_revision_reconcile(self, revision_id):
1078
"""'reconcile' helper routine that allows access to a revision always.
1080
This variant of get_revision does not cross check the weave graph
1081
against the revision one as get_revision does: but it should only
1082
be used by reconcile, or reconcile-alike commands that are correcting
1083
or testing the revision graph.
1085
return self._get_revisions([revision_id])[0]
1088
def get_revisions(self, revision_ids):
1089
"""Get many revisions at once."""
1090
return self._get_revisions(revision_ids)
1093
def _get_revisions(self, revision_ids):
1094
"""Core work logic to get many revisions without sanity checks."""
1095
for rev_id in revision_ids:
1096
if not rev_id or not isinstance(rev_id, basestring):
1097
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1098
revs = self._revision_store.get_revisions(revision_ids,
1099
self.get_transaction())
1103
def get_revision_xml(self, revision_id):
1104
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1105
# would have already do it.
1106
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1107
rev = self.get_revision(revision_id)
1108
rev_tmp = StringIO()
1109
# the current serializer..
1110
self._revision_store._serializer.write_revision(rev, rev_tmp)
1112
return rev_tmp.getvalue()
1114
def get_deltas_for_revisions(self, revisions):
1115
"""Produce a generator of revision deltas.
1117
Note that the input is a sequence of REVISIONS, not revision_ids.
1118
Trees will be held in memory until the generator exits.
1119
Each delta is relative to the revision's lefthand predecessor.
1121
required_trees = set()
1122
for revision in revisions:
1123
required_trees.add(revision.revision_id)
1124
required_trees.update(revision.parent_ids[:1])
1125
trees = dict((t.get_revision_id(), t) for
1126
t in self.revision_trees(required_trees))
1127
for revision in revisions:
1128
if not revision.parent_ids:
1129
old_tree = self.revision_tree(None)
1131
old_tree = trees[revision.parent_ids[0]]
1132
yield trees[revision.revision_id].changes_from(old_tree)
1135
def get_revision_delta(self, revision_id):
1136
"""Return the delta for one revision.
1138
The delta is relative to the left-hand predecessor of the
1141
r = self.get_revision(revision_id)
1142
return list(self.get_deltas_for_revisions([r]))[0]
1145
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1146
signature = gpg_strategy.sign(plaintext)
1147
self.add_signature_text(revision_id, signature)
1150
def add_signature_text(self, revision_id, signature):
1151
self._revision_store.add_revision_signature_text(revision_id,
1153
self.get_transaction())
1155
def find_text_key_references(self):
1156
"""Find the text key references within the repository.
1158
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1159
revision_ids. Each altered file-ids has the exact revision_ids that
1160
altered it listed explicitly.
1161
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1162
to whether they were referred to by the inventory of the
1163
revision_id that they contain. The inventory texts from all present
1164
revision ids are assessed to generate this report.
1166
revision_ids = self.all_revision_ids()
1167
w = self.get_inventory_weave()
1168
pb = ui.ui_factory.nested_progress_bar()
1170
return self._find_text_key_references_from_xml_inventory_lines(
1171
w.iter_lines_added_or_present_in_versions(revision_ids, pb=pb))
1175
def _find_text_key_references_from_xml_inventory_lines(self,
1177
"""Core routine for extracting references to texts from inventories.
1179
This performs the translation of xml lines to revision ids.
1181
:param line_iterator: An iterator of lines, origin_version_id
1182
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1183
to whether they were referred to by the inventory of the
1184
revision_id that they contain. Note that if that revision_id was
1185
not part of the line_iterator's output then False will be given -
1186
even though it may actually refer to that key.
1188
if not self._serializer.support_altered_by_hack:
1189
raise AssertionError(
1190
"_find_text_key_references_from_xml_inventory_lines only "
1191
"supported for branches which store inventory as unnested xml"
1192
", not on %r" % self)
1195
# this code needs to read every new line in every inventory for the
1196
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1197
# not present in one of those inventories is unnecessary but not
1198
# harmful because we are filtering by the revision id marker in the
1199
# inventory lines : we only select file ids altered in one of those
1200
# revisions. We don't need to see all lines in the inventory because
1201
# only those added in an inventory in rev X can contain a revision=X
1203
unescape_revid_cache = {}
1204
unescape_fileid_cache = {}
1206
# jam 20061218 In a big fetch, this handles hundreds of thousands
1207
# of lines, so it has had a lot of inlining and optimizing done.
1208
# Sorry that it is a little bit messy.
1209
# Move several functions to be local variables, since this is a long
1211
search = self._file_ids_altered_regex.search
1212
unescape = _unescape_xml
1213
setdefault = result.setdefault
1214
for line, version_id in line_iterator:
1215
match = search(line)
1218
# One call to match.group() returning multiple items is quite a
1219
# bit faster than 2 calls to match.group() each returning 1
1220
file_id, revision_id = match.group('file_id', 'revision_id')
1222
# Inlining the cache lookups helps a lot when you make 170,000
1223
# lines and 350k ids, versus 8.4 unique ids.
1224
# Using a cache helps in 2 ways:
1225
# 1) Avoids unnecessary decoding calls
1226
# 2) Re-uses cached strings, which helps in future set and
1228
# (2) is enough that removing encoding entirely along with
1229
# the cache (so we are using plain strings) results in no
1230
# performance improvement.
1232
revision_id = unescape_revid_cache[revision_id]
1234
unescaped = unescape(revision_id)
1235
unescape_revid_cache[revision_id] = unescaped
1236
revision_id = unescaped
1238
# Note that unconditionally unescaping means that we deserialise
1239
# every fileid, which for general 'pull' is not great, but we don't
1240
# really want to have some many fulltexts that this matters anyway.
1243
file_id = unescape_fileid_cache[file_id]
1245
unescaped = unescape(file_id)
1246
unescape_fileid_cache[file_id] = unescaped
1249
key = (file_id, revision_id)
1250
setdefault(key, False)
1251
if revision_id == version_id:
1255
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1257
"""Helper routine for fileids_altered_by_revision_ids.
1259
This performs the translation of xml lines to revision ids.
1261
:param line_iterator: An iterator of lines, origin_version_id
1262
:param revision_ids: The revision ids to filter for. This should be a
1263
set or other type which supports efficient __contains__ lookups, as
1264
the revision id from each parsed line will be looked up in the
1265
revision_ids filter.
1266
:return: a dictionary mapping altered file-ids to an iterable of
1267
revision_ids. Each altered file-ids has the exact revision_ids that
1268
altered it listed explicitly.
1271
setdefault = result.setdefault
1272
for file_id, revision_id in \
1273
self._find_text_key_references_from_xml_inventory_lines(
1274
line_iterator).iterkeys():
1275
# once data is all ensured-consistent; then this is
1276
# if revision_id == version_id
1277
if revision_id in revision_ids:
1278
setdefault(file_id, set()).add(revision_id)
1281
def fileids_altered_by_revision_ids(self, revision_ids, _inv_weave=None):
1282
"""Find the file ids and versions affected by revisions.
1284
:param revisions: an iterable containing revision ids.
1285
:param _inv_weave: The inventory weave from this repository or None.
1286
If None, the inventory weave will be opened automatically.
1287
:return: a dictionary mapping altered file-ids to an iterable of
1288
revision_ids. Each altered file-ids has the exact revision_ids that
1289
altered it listed explicitly.
1291
selected_revision_ids = set(revision_ids)
1292
w = _inv_weave or self.get_inventory_weave()
1293
pb = ui.ui_factory.nested_progress_bar()
1295
return self._find_file_ids_from_xml_inventory_lines(
1296
w.iter_lines_added_or_present_in_versions(
1297
selected_revision_ids, pb=pb),
1298
selected_revision_ids)
1302
def iter_files_bytes(self, desired_files):
1303
"""Iterate through file versions.
1305
Files will not necessarily be returned in the order they occur in
1306
desired_files. No specific order is guaranteed.
1308
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1309
value supplied by the caller as part of desired_files. It should
1310
uniquely identify the file version in the caller's context. (Examples:
1311
an index number or a TreeTransform trans_id.)
1313
bytes_iterator is an iterable of bytestrings for the file. The
1314
kind of iterable and length of the bytestrings are unspecified, but for
1315
this implementation, it is a list of lines produced by
1316
VersionedFile.get_lines().
1318
:param desired_files: a list of (file_id, revision_id, identifier)
1321
transaction = self.get_transaction()
1322
for file_id, revision_id, callable_data in desired_files:
1324
weave = self.weave_store.get_weave(file_id, transaction)
1325
except errors.NoSuchFile:
1326
raise errors.NoSuchIdInRepository(self, file_id)
1327
yield callable_data, weave.get_lines(revision_id)
1329
def _generate_text_key_index(self, text_key_references=None,
1331
"""Generate a new text key index for the repository.
1333
This is an expensive function that will take considerable time to run.
1335
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1336
list of parents, also text keys. When a given key has no parents,
1337
the parents list will be [NULL_REVISION].
1339
# All revisions, to find inventory parents.
1340
if ancestors is None:
1341
graph = self.get_graph()
1342
ancestors = graph.get_parent_map(self.all_revision_ids())
1343
if text_key_references is None:
1344
text_key_references = self.find_text_key_references()
1345
pb = ui.ui_factory.nested_progress_bar()
1347
return self._do_generate_text_key_index(ancestors,
1348
text_key_references, pb)
1352
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1353
"""Helper for _generate_text_key_index to avoid deep nesting."""
1354
revision_order = tsort.topo_sort(ancestors)
1355
invalid_keys = set()
1357
for revision_id in revision_order:
1358
revision_keys[revision_id] = set()
1359
text_count = len(text_key_references)
1360
# a cache of the text keys to allow reuse; costs a dict of all the
1361
# keys, but saves a 2-tuple for every child of a given key.
1363
for text_key, valid in text_key_references.iteritems():
1365
invalid_keys.add(text_key)
1367
revision_keys[text_key[1]].add(text_key)
1368
text_key_cache[text_key] = text_key
1369
del text_key_references
1371
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1372
NULL_REVISION = _mod_revision.NULL_REVISION
1373
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1374
# too small for large or very branchy trees. However, for 55K path
1375
# trees, it would be easy to use too much memory trivially. Ideally we
1376
# could gauge this by looking at available real memory etc, but this is
1377
# always a tricky proposition.
1378
inventory_cache = lru_cache.LRUCache(10)
1379
batch_size = 10 # should be ~150MB on a 55K path tree
1380
batch_count = len(revision_order) / batch_size + 1
1382
pb.update("Calculating text parents.", processed_texts, text_count)
1383
for offset in xrange(batch_count):
1384
to_query = revision_order[offset * batch_size:(offset + 1) *
1388
for rev_tree in self.revision_trees(to_query):
1389
revision_id = rev_tree.get_revision_id()
1390
parent_ids = ancestors[revision_id]
1391
for text_key in revision_keys[revision_id]:
1392
pb.update("Calculating text parents.", processed_texts)
1393
processed_texts += 1
1394
candidate_parents = []
1395
for parent_id in parent_ids:
1396
parent_text_key = (text_key[0], parent_id)
1398
check_parent = parent_text_key not in \
1399
revision_keys[parent_id]
1401
# the parent parent_id is a ghost:
1402
check_parent = False
1403
# truncate the derived graph against this ghost.
1404
parent_text_key = None
1406
# look at the parent commit details inventories to
1407
# determine possible candidates in the per file graph.
1410
inv = inventory_cache[parent_id]
1412
inv = self.revision_tree(parent_id).inventory
1413
inventory_cache[parent_id] = inv
1414
parent_entry = inv._byid.get(text_key[0], None)
1415
if parent_entry is not None:
1417
text_key[0], parent_entry.revision)
1419
parent_text_key = None
1420
if parent_text_key is not None:
1421
candidate_parents.append(
1422
text_key_cache[parent_text_key])
1423
parent_heads = text_graph.heads(candidate_parents)
1424
new_parents = list(parent_heads)
1425
new_parents.sort(key=lambda x:candidate_parents.index(x))
1426
if new_parents == []:
1427
new_parents = [NULL_REVISION]
1428
text_index[text_key] = new_parents
1430
for text_key in invalid_keys:
1431
text_index[text_key] = [NULL_REVISION]
1434
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1435
"""Get an iterable listing the keys of all the data introduced by a set
1438
The keys will be ordered so that the corresponding items can be safely
1439
fetched and inserted in that order.
1441
:returns: An iterable producing tuples of (knit-kind, file-id,
1442
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1443
'revisions'. file-id is None unless knit-kind is 'file'.
1445
# XXX: it's a bit weird to control the inventory weave caching in this
1446
# generator. Ideally the caching would be done in fetch.py I think. Or
1447
# maybe this generator should explicitly have the contract that it
1448
# should not be iterated until the previously yielded item has been
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)
1478
yield ("signatures", None, revisions_with_signatures)
1481
yield ("revisions", None, revision_ids)
1484
def get_inventory_weave(self):
1485
return self.control_weaves.get_weave('inventory',
1486
self.get_transaction())
1489
def get_inventory(self, revision_id):
1490
"""Get Inventory object by revision id."""
1491
return self.iter_inventories([revision_id]).next()
1493
def iter_inventories(self, revision_ids):
1494
"""Get many inventories by revision_ids.
1496
This will buffer some or all of the texts used in constructing the
1497
inventories in memory, but will only parse a single inventory at a
1500
:return: An iterator of inventories.
1502
if ((None in revision_ids)
1503
or (_mod_revision.NULL_REVISION in revision_ids)):
1504
raise ValueError('cannot get null revision inventory')
1505
return self._iter_inventories(revision_ids)
1507
def _iter_inventories(self, revision_ids):
1508
"""single-document based inventory iteration."""
1509
texts = self.get_inventory_weave().get_texts(revision_ids)
1510
for text, revision_id in zip(texts, revision_ids):
1511
yield self.deserialise_inventory(revision_id, text)
1513
def deserialise_inventory(self, revision_id, xml):
1514
"""Transform the xml into an inventory object.
1516
:param revision_id: The expected revision id of the inventory.
1517
:param xml: A serialised inventory.
1519
result = self._serializer.read_inventory_from_string(xml, revision_id)
1520
if result.revision_id != revision_id:
1521
raise AssertionError('revision id mismatch %s != %s' % (
1522
result.revision_id, revision_id))
1525
def serialise_inventory(self, inv):
1526
return self._serializer.write_inventory_to_string(inv)
1528
def _serialise_inventory_to_lines(self, inv):
1529
return self._serializer.write_inventory_to_lines(inv)
1531
def get_serializer_format(self):
1532
return self._serializer.format_num
1535
def get_inventory_xml(self, revision_id):
1536
"""Get inventory XML as a file object."""
1538
iw = self.get_inventory_weave()
1539
return iw.get_text(revision_id)
1541
raise errors.HistoryMissing(self, 'inventory', revision_id)
1544
def get_inventory_sha1(self, revision_id):
1545
"""Return the sha1 hash of the inventory entry
1547
return self.get_revision(revision_id).inventory_sha1
1549
def iter_reverse_revision_history(self, revision_id):
1550
"""Iterate backwards through revision ids in the lefthand history
1552
:param revision_id: The revision id to start with. All its lefthand
1553
ancestors will be traversed.
1555
graph = self.get_graph()
1556
next_id = revision_id
1558
if next_id in (None, _mod_revision.NULL_REVISION):
1561
# Note: The following line may raise KeyError in the event of
1562
# truncated history. We decided not to have a try:except:raise
1563
# RevisionNotPresent here until we see a use for it, because of the
1564
# cost in an inner loop that is by its very nature O(history).
1565
# Robert Collins 20080326
1566
parents = graph.get_parent_map([next_id])[next_id]
1567
if len(parents) == 0:
1570
next_id = parents[0]
1573
def get_revision_inventory(self, revision_id):
1574
"""Return inventory of a past revision."""
1575
# TODO: Unify this with get_inventory()
1576
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1577
# must be the same as its revision, so this is trivial.
1578
if revision_id is None:
1579
# This does not make sense: if there is no revision,
1580
# then it is the current tree inventory surely ?!
1581
# and thus get_root_id() is something that looks at the last
1582
# commit on the branch, and the get_root_id is an inventory check.
1583
raise NotImplementedError
1584
# return Inventory(self.get_root_id())
1586
return self.get_inventory(revision_id)
1589
def is_shared(self):
1590
"""Return True if this repository is flagged as a shared repository."""
1591
raise NotImplementedError(self.is_shared)
1594
def reconcile(self, other=None, thorough=False):
1595
"""Reconcile this repository."""
1596
from bzrlib.reconcile import RepoReconciler
1597
reconciler = RepoReconciler(self, thorough=thorough)
1598
reconciler.reconcile()
1601
def _refresh_data(self):
1602
"""Helper called from lock_* to ensure coherency with disk.
1604
The default implementation does nothing; it is however possible
1605
for repositories to maintain loaded indices across multiple locks
1606
by checking inside their implementation of this method to see
1607
whether their indices are still valid. This depends of course on
1608
the disk format being validatable in this manner.
1612
def revision_tree(self, revision_id):
1613
"""Return Tree for a revision on this branch.
1615
`revision_id` may be None for the empty tree revision.
1617
# TODO: refactor this to use an existing revision object
1618
# so we don't need to read it in twice.
1619
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1620
return RevisionTree(self, Inventory(root_id=None),
1621
_mod_revision.NULL_REVISION)
1623
inv = self.get_revision_inventory(revision_id)
1624
return RevisionTree(self, inv, revision_id)
1626
def revision_trees(self, revision_ids):
1627
"""Return Tree for a revision on this branch.
1629
`revision_id` may not be None or 'null:'"""
1630
inventories = self.iter_inventories(revision_ids)
1631
for inv in inventories:
1632
yield RevisionTree(self, inv, inv.revision_id)
1635
def get_ancestry(self, revision_id, topo_sorted=True):
1636
"""Return a list of revision-ids integrated by a revision.
1638
The first element of the list is always None, indicating the origin
1639
revision. This might change when we have history horizons, or
1640
perhaps we should have a new API.
1642
This is topologically sorted.
1644
if _mod_revision.is_null(revision_id):
1646
if not self.has_revision(revision_id):
1647
raise errors.NoSuchRevision(self, revision_id)
1648
w = self.get_inventory_weave()
1649
candidates = w.get_ancestry(revision_id, topo_sorted)
1650
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1653
"""Compress the data within the repository.
1655
This operation only makes sense for some repository types. For other
1656
types it should be a no-op that just returns.
1658
This stub method does not require a lock, but subclasses should use
1659
@needs_write_lock as this is a long running call its reasonable to
1660
implicitly lock for the user.
1664
def print_file(self, file, revision_id):
1665
"""Print `file` to stdout.
1667
FIXME RBC 20060125 as John Meinel points out this is a bad api
1668
- it writes to stdout, it assumes that that is valid etc. Fix
1669
by creating a new more flexible convenience function.
1671
tree = self.revision_tree(revision_id)
1672
# use inventory as it was in that revision
1673
file_id = tree.inventory.path2id(file)
1675
# TODO: jam 20060427 Write a test for this code path
1676
# it had a bug in it, and was raising the wrong
1678
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1679
tree.print_file(file_id)
1681
def get_transaction(self):
1682
return self.control_files.get_transaction()
1684
@deprecated_method(symbol_versioning.one_five)
1685
def revision_parents(self, revision_id):
1686
return self.get_inventory_weave().parent_names(revision_id)
1688
@deprecated_method(symbol_versioning.one_one)
1689
def get_parents(self, revision_ids):
1690
"""See StackedParentsProvider.get_parents"""
1691
parent_map = self.get_parent_map(revision_ids)
1692
return [parent_map.get(r, None) for r in revision_ids]
1694
def get_parent_map(self, keys):
1695
"""See graph._StackedParentsProvider.get_parent_map"""
1697
for revision_id in keys:
1698
if revision_id is None:
1699
raise ValueError('get_parent_map(None) is not valid')
1700
if revision_id == _mod_revision.NULL_REVISION:
1701
parent_map[revision_id] = ()
1704
parent_id_list = self.get_revision(revision_id).parent_ids
1705
except errors.NoSuchRevision:
1708
if len(parent_id_list) == 0:
1709
parent_ids = (_mod_revision.NULL_REVISION,)
1711
parent_ids = tuple(parent_id_list)
1712
parent_map[revision_id] = parent_ids
1715
def _make_parents_provider(self):
1718
def get_graph(self, other_repository=None):
1719
"""Return the graph walker for this repository format"""
1720
parents_provider = self._make_parents_provider()
1721
if (other_repository is not None and
1722
not self.has_same_location(other_repository)):
1723
parents_provider = graph._StackedParentsProvider(
1724
[parents_provider, other_repository._make_parents_provider()])
1725
return graph.Graph(parents_provider)
1727
def _get_versioned_file_checker(self):
1728
"""Return an object suitable for checking versioned files."""
1729
return _VersionedFileChecker(self)
1731
def revision_ids_to_search_result(self, result_set):
1732
"""Convert a set of revision ids to a graph SearchResult."""
1733
result_parents = set()
1734
for parents in self.get_graph().get_parent_map(
1735
result_set).itervalues():
1736
result_parents.update(parents)
1737
included_keys = result_set.intersection(result_parents)
1738
start_keys = result_set.difference(included_keys)
1739
exclude_keys = result_parents.difference(result_set)
1740
result = graph.SearchResult(start_keys, exclude_keys,
1741
len(result_set), result_set)
1745
def set_make_working_trees(self, new_value):
1746
"""Set the policy flag for making working trees when creating branches.
1748
This only applies to branches that use this repository.
1750
The default is 'True'.
1751
:param new_value: True to restore the default, False to disable making
1754
raise NotImplementedError(self.set_make_working_trees)
1756
def make_working_trees(self):
1757
"""Returns the policy for making working trees on new branches."""
1758
raise NotImplementedError(self.make_working_trees)
1761
def sign_revision(self, revision_id, gpg_strategy):
1762
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1763
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1766
def has_signature_for_revision_id(self, revision_id):
1767
"""Query for a revision signature for revision_id in the repository."""
1768
return self._revision_store.has_signature(revision_id,
1769
self.get_transaction())
1772
def get_signature_text(self, revision_id):
1773
"""Return the text for a signature."""
1774
return self._revision_store.get_signature_text(revision_id,
1775
self.get_transaction())
1778
def check(self, revision_ids=None):
1779
"""Check consistency of all history of given revision_ids.
1781
Different repository implementations should override _check().
1783
:param revision_ids: A non-empty list of revision_ids whose ancestry
1784
will be checked. Typically the last revision_id of a branch.
1786
return self._check(revision_ids)
1788
def _check(self, revision_ids):
1789
result = check.Check(self)
1793
def _warn_if_deprecated(self):
1794
global _deprecation_warning_done
1795
if _deprecation_warning_done:
1797
_deprecation_warning_done = True
1798
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1799
% (self._format, self.bzrdir.transport.base))
1801
def supports_rich_root(self):
1802
return self._format.rich_root_data
1804
def _check_ascii_revisionid(self, revision_id, method):
1805
"""Private helper for ascii-only repositories."""
1806
# weave repositories refuse to store revisionids that are non-ascii.
1807
if revision_id is not None:
1808
# weaves require ascii revision ids.
1809
if isinstance(revision_id, unicode):
1811
revision_id.encode('ascii')
1812
except UnicodeEncodeError:
1813
raise errors.NonAsciiRevisionId(method, self)
1816
revision_id.decode('ascii')
1817
except UnicodeDecodeError:
1818
raise errors.NonAsciiRevisionId(method, self)
1820
def revision_graph_can_have_wrong_parents(self):
1821
"""Is it possible for this repository to have a revision graph with
1824
If True, then this repository must also implement
1825
_find_inconsistent_revision_parents so that check and reconcile can
1826
check for inconsistencies before proceeding with other checks that may
1827
depend on the revision index being consistent.
1829
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1832
# remove these delegates a while after bzr 0.15
1833
def __make_delegated(name, from_module):
1834
def _deprecated_repository_forwarder():
1835
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1836
% (name, from_module),
1839
m = __import__(from_module, globals(), locals(), [name])
1841
return getattr(m, name)
1842
except AttributeError:
1843
raise AttributeError('module %s has no name %s'
1845
globals()[name] = _deprecated_repository_forwarder
1848
'AllInOneRepository',
1849
'WeaveMetaDirRepository',
1850
'PreSplitOutRepositoryFormat',
1851
'RepositoryFormat4',
1852
'RepositoryFormat5',
1853
'RepositoryFormat6',
1854
'RepositoryFormat7',
1856
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1860
'RepositoryFormatKnit',
1861
'RepositoryFormatKnit1',
1863
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1866
def install_revision(repository, rev, revision_tree):
1867
"""Install all revision data into a repository."""
1868
install_revisions(repository, [(rev, revision_tree, None)])
1871
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1872
"""Install all revision data into a repository.
1874
Accepts an iterable of revision, tree, signature tuples. The signature
1877
repository.start_write_group()
1879
for n, (revision, revision_tree, signature) in enumerate(iterable):
1880
_install_revision(repository, revision, revision_tree, signature)
1882
pb.update('Transferring revisions', n + 1, num_revisions)
1884
repository.abort_write_group()
1887
repository.commit_write_group()
1890
def _install_revision(repository, rev, revision_tree, signature):
1891
"""Install all revision data into a repository."""
1892
present_parents = []
1894
for p_id in rev.parent_ids:
1895
if repository.has_revision(p_id):
1896
present_parents.append(p_id)
1897
parent_trees[p_id] = repository.revision_tree(p_id)
1899
parent_trees[p_id] = repository.revision_tree(None)
1901
inv = revision_tree.inventory
1902
entries = inv.iter_entries()
1903
# backwards compatibility hack: skip the root id.
1904
if not repository.supports_rich_root():
1905
path, root = entries.next()
1906
if root.revision != rev.revision_id:
1907
raise errors.IncompatibleRevision(repr(repository))
1908
# Add the texts that are not already present
1909
for path, ie in entries:
1910
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1911
repository.get_transaction())
1912
if ie.revision not in w:
1914
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1915
# with InventoryEntry.find_previous_heads(). if it is, then there
1916
# is a latent bug here where the parents may have ancestors of each
1918
for revision, tree in parent_trees.iteritems():
1919
if ie.file_id not in tree:
1921
parent_id = tree.inventory[ie.file_id].revision
1922
if parent_id in text_parents:
1924
text_parents.append(parent_id)
1926
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1927
repository.get_transaction())
1928
lines = revision_tree.get_file(ie.file_id).readlines()
1929
vfile.add_lines(rev.revision_id, text_parents, lines)
1931
# install the inventory
1932
repository.add_inventory(rev.revision_id, inv, present_parents)
1933
except errors.RevisionAlreadyPresent:
1935
if signature is not None:
1936
repository.add_signature_text(rev.revision_id, signature)
1937
repository.add_revision(rev.revision_id, rev, inv)
1940
class MetaDirRepository(Repository):
1941
"""Repositories in the new meta-dir layout."""
1943
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1944
super(MetaDirRepository, self).__init__(_format,
1950
dir_mode = self.control_files._dir_mode
1951
file_mode = self.control_files._file_mode
1954
def is_shared(self):
1955
"""Return True if this repository is flagged as a shared repository."""
1956
return self.control_files._transport.has('shared-storage')
1959
def set_make_working_trees(self, new_value):
1960
"""Set the policy flag for making working trees when creating branches.
1962
This only applies to branches that use this repository.
1964
The default is 'True'.
1965
:param new_value: True to restore the default, False to disable making
1970
self.control_files._transport.delete('no-working-trees')
1971
except errors.NoSuchFile:
1974
self.control_files.put_utf8('no-working-trees', '')
1976
def make_working_trees(self):
1977
"""Returns the policy for making working trees on new branches."""
1978
return not self.control_files._transport.has('no-working-trees')
1981
class MetaDirVersionedFileRepository(MetaDirRepository):
1982
"""Repositories in a meta-dir, that work via versioned file objects."""
1984
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1985
super(MetaDirVersionedFileRepository, self).__init__(_format, a_bzrdir,
1986
control_files, _revision_store, control_store, text_store)
1987
_revision_store.get_scope = self.get_transaction
1988
control_store.get_scope = self.get_transaction
1989
text_store.get_scope = self.get_transaction
1992
class RepositoryFormatRegistry(registry.Registry):
1993
"""Registry of RepositoryFormats."""
1995
def get(self, format_string):
1996
r = registry.Registry.get(self, format_string)
2002
format_registry = RepositoryFormatRegistry()
2003
"""Registry of formats, indexed by their identifying format string.
2005
This can contain either format instances themselves, or classes/factories that
2006
can be called to obtain one.
2010
#####################################################################
2011
# Repository Formats
2013
class RepositoryFormat(object):
2014
"""A repository format.
2016
Formats provide three things:
2017
* An initialization routine to construct repository data on disk.
2018
* a format string which is used when the BzrDir supports versioned
2020
* an open routine which returns a Repository instance.
2022
There is one and only one Format subclass for each on-disk format. But
2023
there can be one Repository subclass that is used for several different
2024
formats. The _format attribute on a Repository instance can be used to
2025
determine the disk format.
2027
Formats are placed in an dict by their format string for reference
2028
during opening. These should be subclasses of RepositoryFormat
2031
Once a format is deprecated, just deprecate the initialize and open
2032
methods on the format class. Do not deprecate the object, as the
2033
object will be created every system load.
2035
Common instance attributes:
2036
_matchingbzrdir - the bzrdir format that the repository format was
2037
originally written to work with. This can be used if manually
2038
constructing a bzrdir and repository, or more commonly for test suite
2042
# Set to True or False in derived classes. True indicates that the format
2043
# supports ghosts gracefully.
2044
supports_ghosts = None
2045
# Can this repository be given external locations to lookup additional
2046
# data. Set to True or False in derived classes.
2047
supports_external_lookups = None
2050
return "<%s>" % self.__class__.__name__
2052
def __eq__(self, other):
2053
# format objects are generally stateless
2054
return isinstance(other, self.__class__)
2056
def __ne__(self, other):
2057
return not self == other
2060
def find_format(klass, a_bzrdir):
2061
"""Return the format for the repository object in a_bzrdir.
2063
This is used by bzr native formats that have a "format" file in
2064
the repository. Other methods may be used by different types of
2068
transport = a_bzrdir.get_repository_transport(None)
2069
format_string = transport.get("format").read()
2070
return format_registry.get(format_string)
2071
except errors.NoSuchFile:
2072
raise errors.NoRepositoryPresent(a_bzrdir)
2074
raise errors.UnknownFormatError(format=format_string,
2078
def register_format(klass, format):
2079
format_registry.register(format.get_format_string(), format)
2082
def unregister_format(klass, format):
2083
format_registry.remove(format.get_format_string())
2086
def get_default_format(klass):
2087
"""Return the current default format."""
2088
from bzrlib import bzrdir
2089
return bzrdir.format_registry.make_bzrdir('default').repository_format
2091
def _get_control_store(self, repo_transport, control_files):
2092
"""Return the control store for this repository."""
2093
raise NotImplementedError(self._get_control_store)
2095
def get_format_string(self):
2096
"""Return the ASCII format string that identifies this format.
2098
Note that in pre format ?? repositories the format string is
2099
not permitted nor written to disk.
2101
raise NotImplementedError(self.get_format_string)
2103
def get_format_description(self):
2104
"""Return the short description for this format."""
2105
raise NotImplementedError(self.get_format_description)
2107
def _get_revision_store(self, repo_transport, control_files):
2108
"""Return the revision store object for this a_bzrdir."""
2109
raise NotImplementedError(self._get_revision_store)
2111
def _get_text_rev_store(self,
2118
"""Common logic for getting a revision store for a repository.
2120
see self._get_revision_store for the subclass-overridable method to
2121
get the store for a repository.
2123
from bzrlib.store.revision.text import TextRevisionStore
2124
dir_mode = control_files._dir_mode
2125
file_mode = control_files._file_mode
2126
text_store = TextStore(transport.clone(name),
2128
compressed=compressed,
2130
file_mode=file_mode)
2131
_revision_store = TextRevisionStore(text_store, serializer)
2132
return _revision_store
2134
# TODO: this shouldn't be in the base class, it's specific to things that
2135
# use weaves or knits -- mbp 20070207
2136
def _get_versioned_file_store(self,
2141
versionedfile_class=None,
2142
versionedfile_kwargs={},
2144
if versionedfile_class is None:
2145
versionedfile_class = self._versionedfile_class
2146
weave_transport = control_files._transport.clone(name)
2147
dir_mode = control_files._dir_mode
2148
file_mode = control_files._file_mode
2149
return VersionedFileStore(weave_transport, prefixed=prefixed,
2151
file_mode=file_mode,
2152
versionedfile_class=versionedfile_class,
2153
versionedfile_kwargs=versionedfile_kwargs,
2156
def initialize(self, a_bzrdir, shared=False):
2157
"""Initialize a repository of this format in a_bzrdir.
2159
:param a_bzrdir: The bzrdir to put the new repository in it.
2160
:param shared: The repository should be initialized as a sharable one.
2161
:returns: The new repository object.
2163
This may raise UninitializableFormat if shared repository are not
2164
compatible the a_bzrdir.
2166
raise NotImplementedError(self.initialize)
2168
def is_supported(self):
2169
"""Is this format supported?
2171
Supported formats must be initializable and openable.
2172
Unsupported formats may not support initialization or committing or
2173
some other features depending on the reason for not being supported.
2177
def check_conversion_target(self, target_format):
2178
raise NotImplementedError(self.check_conversion_target)
2180
def open(self, a_bzrdir, _found=False):
2181
"""Return an instance of this format for the bzrdir a_bzrdir.
2183
_found is a private parameter, do not use it.
2185
raise NotImplementedError(self.open)
2188
class MetaDirRepositoryFormat(RepositoryFormat):
2189
"""Common base class for the new repositories using the metadir layout."""
2191
rich_root_data = False
2192
supports_tree_reference = False
2193
supports_external_lookups = False
2194
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
2197
super(MetaDirRepositoryFormat, self).__init__()
2199
def _create_control_files(self, a_bzrdir):
2200
"""Create the required files and the initial control_files object."""
2201
# FIXME: RBC 20060125 don't peek under the covers
2202
# NB: no need to escape relative paths that are url safe.
2203
repository_transport = a_bzrdir.get_repository_transport(self)
2204
control_files = lockable_files.LockableFiles(repository_transport,
2205
'lock', lockdir.LockDir)
2206
control_files.create_lock()
2207
return control_files
2209
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
2210
"""Upload the initial blank content."""
2211
control_files = self._create_control_files(a_bzrdir)
2212
control_files.lock_write()
2214
control_files._transport.mkdir_multi(dirs,
2215
mode=control_files._dir_mode)
2216
for file, content in files:
2217
control_files.put(file, content)
2218
for file, content in utf8_files:
2219
control_files.put_utf8(file, content)
2221
control_files.put_utf8('shared-storage', '')
2223
control_files.unlock()
2226
# formats which have no format string are not discoverable
2227
# and not independently creatable, so are not registered. They're
2228
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
2229
# needed, it's constructed directly by the BzrDir. Non-native formats where
2230
# the repository is not separately opened are similar.
2232
format_registry.register_lazy(
2233
'Bazaar-NG Repository format 7',
2234
'bzrlib.repofmt.weaverepo',
2238
format_registry.register_lazy(
2239
'Bazaar-NG Knit Repository Format 1',
2240
'bzrlib.repofmt.knitrepo',
2241
'RepositoryFormatKnit1',
2244
format_registry.register_lazy(
2245
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
2246
'bzrlib.repofmt.knitrepo',
2247
'RepositoryFormatKnit3',
2250
format_registry.register_lazy(
2251
'Bazaar Knit Repository Format 4 (bzr 1.0)\n',
2252
'bzrlib.repofmt.knitrepo',
2253
'RepositoryFormatKnit4',
2256
# Pack-based formats. There is one format for pre-subtrees, and one for
2257
# post-subtrees to allow ease of testing.
2258
# NOTE: These are experimental in 0.92. Stable in 1.0 and above
2259
format_registry.register_lazy(
2260
'Bazaar pack repository format 1 (needs bzr 0.92)\n',
2261
'bzrlib.repofmt.pack_repo',
2262
'RepositoryFormatKnitPack1',
2264
format_registry.register_lazy(
2265
'Bazaar pack repository format 1 with subtree support (needs bzr 0.92)\n',
2266
'bzrlib.repofmt.pack_repo',
2267
'RepositoryFormatKnitPack3',
2269
format_registry.register_lazy(
2270
'Bazaar pack repository format 1 with rich root (needs bzr 1.0)\n',
2271
'bzrlib.repofmt.pack_repo',
2272
'RepositoryFormatKnitPack4',
2274
# Development formats.
2276
# development 0 - stub to introduce development versioning scheme.
2277
format_registry.register_lazy(
2278
"Bazaar development format 0 (needs bzr.dev from before 1.3)\n",
2279
'bzrlib.repofmt.pack_repo',
2280
'RepositoryFormatPackDevelopment0',
2282
format_registry.register_lazy(
2283
("Bazaar development format 0 with subtree support "
2284
"(needs bzr.dev from before 1.3)\n"),
2285
'bzrlib.repofmt.pack_repo',
2286
'RepositoryFormatPackDevelopment0Subtree',
2288
# 1.3->1.4 go below here
2291
class InterRepository(InterObject):
2292
"""This class represents operations taking place between two repositories.
2294
Its instances have methods like copy_content and fetch, and contain
2295
references to the source and target repositories these operations can be
2298
Often we will provide convenience methods on 'repository' which carry out
2299
operations with another repository - they will always forward to
2300
InterRepository.get(other).method_name(parameters).
2304
"""The available optimised InterRepository types."""
2306
def copy_content(self, revision_id=None):
2307
raise NotImplementedError(self.copy_content)
2309
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2310
"""Fetch the content required to construct revision_id.
2312
The content is copied from self.source to self.target.
2314
:param revision_id: if None all content is copied, if NULL_REVISION no
2316
:param pb: optional progress bar to use for progress reports. If not
2317
provided a default one will be created.
2319
Returns the copied revision count and the failed revisions in a tuple:
2322
raise NotImplementedError(self.fetch)
2324
def _walk_to_common_revisions(self, revision_ids):
2325
"""Walk out from revision_ids in source to revisions target has.
2327
:param revision_ids: The start point for the search.
2328
:return: A set of revision ids.
2330
target_graph = self.target.get_graph()
2331
revision_ids = frozenset(revision_ids)
2332
if set(target_graph.get_parent_map(revision_ids)) == revision_ids:
2333
return graph.SearchResult(revision_ids, set(), 0, set())
2334
missing_revs = set()
2335
source_graph = self.source.get_graph()
2336
# ensure we don't pay silly lookup costs.
2337
searcher = source_graph._make_breadth_first_searcher(revision_ids)
2338
null_set = frozenset([_mod_revision.NULL_REVISION])
2341
next_revs, ghosts = searcher.next_with_ghosts()
2342
except StopIteration:
2344
if revision_ids.intersection(ghosts):
2345
absent_ids = set(revision_ids.intersection(ghosts))
2346
# If all absent_ids are present in target, no error is needed.
2347
absent_ids.difference_update(
2348
set(target_graph.get_parent_map(absent_ids)))
2350
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2351
# we don't care about other ghosts as we can't fetch them and
2352
# haven't been asked to.
2353
next_revs = set(next_revs)
2354
# we always have NULL_REVISION present.
2355
have_revs = set(target_graph.get_parent_map(next_revs)).union(null_set)
2356
missing_revs.update(next_revs - have_revs)
2357
searcher.stop_searching_any(have_revs)
2358
return searcher.get_result()
2360
@deprecated_method(symbol_versioning.one_two)
2362
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2363
"""Return the revision ids that source has that target does not.
2365
These are returned in topological order.
2367
:param revision_id: only return revision ids included by this
2369
:param find_ghosts: If True find missing revisions in deep history
2370
rather than just finding the surface difference.
2372
return list(self.search_missing_revision_ids(
2373
revision_id, find_ghosts).get_keys())
2376
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2377
"""Return the revision ids that source has that target does not.
2379
:param revision_id: only return revision ids included by this
2381
:param find_ghosts: If True find missing revisions in deep history
2382
rather than just finding the surface difference.
2383
:return: A bzrlib.graph.SearchResult.
2385
# stop searching at found target revisions.
2386
if not find_ghosts and revision_id is not None:
2387
return self._walk_to_common_revisions([revision_id])
2388
# generic, possibly worst case, slow code path.
2389
target_ids = set(self.target.all_revision_ids())
2390
if revision_id is not None:
2391
source_ids = self.source.get_ancestry(revision_id)
2392
if source_ids[0] is not None:
2393
raise AssertionError()
2396
source_ids = self.source.all_revision_ids()
2397
result_set = set(source_ids).difference(target_ids)
2398
return self.source.revision_ids_to_search_result(result_set)
2401
def _same_model(source, target):
2402
"""True if source and target have the same data representation."""
2403
if source.supports_rich_root() != target.supports_rich_root():
2405
if source._serializer != target._serializer:
2410
class InterSameDataRepository(InterRepository):
2411
"""Code for converting between repositories that represent the same data.
2413
Data format and model must match for this to work.
2417
def _get_repo_format_to_test(self):
2418
"""Repository format for testing with.
2420
InterSameData can pull from subtree to subtree and from non-subtree to
2421
non-subtree, so we test this with the richest repository format.
2423
from bzrlib.repofmt import knitrepo
2424
return knitrepo.RepositoryFormatKnit3()
2427
def is_compatible(source, target):
2428
return InterRepository._same_model(source, target)
2431
def copy_content(self, revision_id=None):
2432
"""Make a complete copy of the content in self into destination.
2434
This copies both the repository's revision data, and configuration information
2435
such as the make_working_trees setting.
2437
This is a destructive operation! Do not use it on existing
2440
:param revision_id: Only copy the content needed to construct
2441
revision_id and its parents.
2444
self.target.set_make_working_trees(self.source.make_working_trees())
2445
except NotImplementedError:
2447
# but don't bother fetching if we have the needed data now.
2448
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2449
self.target.has_revision(revision_id)):
2451
self.target.fetch(self.source, revision_id=revision_id)
2454
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2455
"""See InterRepository.fetch()."""
2456
from bzrlib.fetch import GenericRepoFetcher
2457
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2458
self.source, self.source._format, self.target,
2459
self.target._format)
2460
f = GenericRepoFetcher(to_repository=self.target,
2461
from_repository=self.source,
2462
last_revision=revision_id,
2463
pb=pb, find_ghosts=find_ghosts)
2464
return f.count_copied, f.failed_revisions
2467
class InterWeaveRepo(InterSameDataRepository):
2468
"""Optimised code paths between Weave based repositories.
2470
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2471
implemented lazy inter-object optimisation.
2475
def _get_repo_format_to_test(self):
2476
from bzrlib.repofmt import weaverepo
2477
return weaverepo.RepositoryFormat7()
2480
def is_compatible(source, target):
2481
"""Be compatible with known Weave formats.
2483
We don't test for the stores being of specific types because that
2484
could lead to confusing results, and there is no need to be
2487
from bzrlib.repofmt.weaverepo import (
2493
return (isinstance(source._format, (RepositoryFormat5,
2495
RepositoryFormat7)) and
2496
isinstance(target._format, (RepositoryFormat5,
2498
RepositoryFormat7)))
2499
except AttributeError:
2503
def copy_content(self, revision_id=None):
2504
"""See InterRepository.copy_content()."""
2505
# weave specific optimised path:
2507
self.target.set_make_working_trees(self.source.make_working_trees())
2508
except (errors.RepositoryUpgradeRequired, NotImplemented):
2510
# FIXME do not peek!
2511
if self.source.control_files._transport.listable():
2512
pb = ui.ui_factory.nested_progress_bar()
2514
self.target.weave_store.copy_all_ids(
2515
self.source.weave_store,
2517
from_transaction=self.source.get_transaction(),
2518
to_transaction=self.target.get_transaction())
2519
pb.update('copying inventory', 0, 1)
2520
self.target.control_weaves.copy_multi(
2521
self.source.control_weaves, ['inventory'],
2522
from_transaction=self.source.get_transaction(),
2523
to_transaction=self.target.get_transaction())
2524
self.target._revision_store.text_store.copy_all_ids(
2525
self.source._revision_store.text_store,
2530
self.target.fetch(self.source, revision_id=revision_id)
2533
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2534
"""See InterRepository.fetch()."""
2535
from bzrlib.fetch import GenericRepoFetcher
2536
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2537
self.source, self.source._format, self.target, self.target._format)
2538
f = GenericRepoFetcher(to_repository=self.target,
2539
from_repository=self.source,
2540
last_revision=revision_id,
2541
pb=pb, find_ghosts=find_ghosts)
2542
return f.count_copied, f.failed_revisions
2545
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2546
"""See InterRepository.missing_revision_ids()."""
2547
# we want all revisions to satisfy revision_id in source.
2548
# but we don't want to stat every file here and there.
2549
# we want then, all revisions other needs to satisfy revision_id
2550
# checked, but not those that we have locally.
2551
# so the first thing is to get a subset of the revisions to
2552
# satisfy revision_id in source, and then eliminate those that
2553
# we do already have.
2554
# this is slow on high latency connection to self, but as as this
2555
# disk format scales terribly for push anyway due to rewriting
2556
# inventory.weave, this is considered acceptable.
2558
if revision_id is not None:
2559
source_ids = self.source.get_ancestry(revision_id)
2560
if source_ids[0] is not None:
2561
raise AssertionError()
2564
source_ids = self.source._all_possible_ids()
2565
source_ids_set = set(source_ids)
2566
# source_ids is the worst possible case we may need to pull.
2567
# now we want to filter source_ids against what we actually
2568
# have in target, but don't try to check for existence where we know
2569
# we do not have a revision as that would be pointless.
2570
target_ids = set(self.target._all_possible_ids())
2571
possibly_present_revisions = target_ids.intersection(source_ids_set)
2572
actually_present_revisions = set(
2573
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2574
required_revisions = source_ids_set.difference(actually_present_revisions)
2575
if revision_id is not None:
2576
# we used get_ancestry to determine source_ids then we are assured all
2577
# revisions referenced are present as they are installed in topological order.
2578
# and the tip revision was validated by get_ancestry.
2579
result_set = required_revisions
2581
# if we just grabbed the possibly available ids, then
2582
# we only have an estimate of whats available and need to validate
2583
# that against the revision records.
2585
self.source._eliminate_revisions_not_present(required_revisions))
2586
return self.source.revision_ids_to_search_result(result_set)
2589
class InterKnitRepo(InterSameDataRepository):
2590
"""Optimised code paths between Knit based repositories."""
2593
def _get_repo_format_to_test(self):
2594
from bzrlib.repofmt import knitrepo
2595
return knitrepo.RepositoryFormatKnit1()
2598
def is_compatible(source, target):
2599
"""Be compatible with known Knit formats.
2601
We don't test for the stores being of specific types because that
2602
could lead to confusing results, and there is no need to be
2605
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2607
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2608
isinstance(target._format, RepositoryFormatKnit))
2609
except AttributeError:
2611
return are_knits and InterRepository._same_model(source, target)
2614
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2615
"""See InterRepository.fetch()."""
2616
from bzrlib.fetch import KnitRepoFetcher
2617
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2618
self.source, self.source._format, self.target, self.target._format)
2619
f = KnitRepoFetcher(to_repository=self.target,
2620
from_repository=self.source,
2621
last_revision=revision_id,
2622
pb=pb, find_ghosts=find_ghosts)
2623
return f.count_copied, f.failed_revisions
2626
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2627
"""See InterRepository.missing_revision_ids()."""
2628
if revision_id is not None:
2629
source_ids = self.source.get_ancestry(revision_id)
2630
if source_ids[0] is not None:
2631
raise AssertionError()
2634
source_ids = self.source.all_revision_ids()
2635
source_ids_set = set(source_ids)
2636
# source_ids is the worst possible case we may need to pull.
2637
# now we want to filter source_ids against what we actually
2638
# have in target, but don't try to check for existence where we know
2639
# we do not have a revision as that would be pointless.
2640
target_ids = set(self.target.all_revision_ids())
2641
possibly_present_revisions = target_ids.intersection(source_ids_set)
2642
actually_present_revisions = set(
2643
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2644
required_revisions = source_ids_set.difference(actually_present_revisions)
2645
if revision_id is not None:
2646
# we used get_ancestry to determine source_ids then we are assured all
2647
# revisions referenced are present as they are installed in topological order.
2648
# and the tip revision was validated by get_ancestry.
2649
result_set = required_revisions
2651
# if we just grabbed the possibly available ids, then
2652
# we only have an estimate of whats available and need to validate
2653
# that against the revision records.
2655
self.source._eliminate_revisions_not_present(required_revisions))
2656
return self.source.revision_ids_to_search_result(result_set)
2659
class InterPackRepo(InterSameDataRepository):
2660
"""Optimised code paths between Pack based repositories."""
2663
def _get_repo_format_to_test(self):
2664
from bzrlib.repofmt import pack_repo
2665
return pack_repo.RepositoryFormatKnitPack1()
2668
def is_compatible(source, target):
2669
"""Be compatible with known Pack formats.
2671
We don't test for the stores being of specific types because that
2672
could lead to confusing results, and there is no need to be
2675
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2677
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2678
isinstance(target._format, RepositoryFormatPack))
2679
except AttributeError:
2681
return are_packs and InterRepository._same_model(source, target)
2684
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2685
"""See InterRepository.fetch()."""
2686
from bzrlib.repofmt.pack_repo import Packer
2687
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2688
self.source, self.source._format, self.target, self.target._format)
2689
self.count_copied = 0
2690
if revision_id is None:
2692
# everything to do - use pack logic
2693
# to fetch from all packs to one without
2694
# inventory parsing etc, IFF nothing to be copied is in the target.
2696
revision_ids = self.source.all_revision_ids()
2697
revision_keys = [(revid,) for revid in revision_ids]
2698
index = self.target._pack_collection.revision_index.combined_index
2699
present_revision_ids = set(item[1][0] for item in
2700
index.iter_entries(revision_keys))
2701
revision_ids = set(revision_ids) - present_revision_ids
2702
# implementing the TODO will involve:
2703
# - detecting when all of a pack is selected
2704
# - avoiding as much as possible pre-selection, so the
2705
# more-core routines such as create_pack_from_packs can filter in
2706
# a just-in-time fashion. (though having a HEADS list on a
2707
# repository might make this a lot easier, because we could
2708
# sensibly detect 'new revisions' without doing a full index scan.
2709
elif _mod_revision.is_null(revision_id):
2714
revision_ids = self.search_missing_revision_ids(revision_id,
2715
find_ghosts=find_ghosts).get_keys()
2716
except errors.NoSuchRevision:
2717
raise errors.InstallFailed([revision_id])
2718
if len(revision_ids) == 0:
2720
packs = self.source._pack_collection.all_packs()
2721
pack = Packer(self.target._pack_collection, packs, '.fetch',
2722
revision_ids).pack()
2723
if pack is not None:
2724
self.target._pack_collection._save_pack_names()
2725
# Trigger an autopack. This may duplicate effort as we've just done
2726
# a pack creation, but for now it is simpler to think about as
2727
# 'upload data, then repack if needed'.
2728
self.target._pack_collection.autopack()
2729
return (pack.get_revision_count(), [])
2734
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2735
"""See InterRepository.missing_revision_ids().
2737
:param find_ghosts: Find ghosts throughout the ancestry of
2740
if not find_ghosts and revision_id is not None:
2741
return self._walk_to_common_revisions([revision_id])
2742
elif revision_id is not None:
2743
source_ids = self.source.get_ancestry(revision_id)
2744
if source_ids[0] is not None:
2745
raise AssertionError()
2748
source_ids = self.source.all_revision_ids()
2749
# source_ids is the worst possible case we may need to pull.
2750
# now we want to filter source_ids against what we actually
2751
# have in target, but don't try to check for existence where we know
2752
# we do not have a revision as that would be pointless.
2753
target_ids = set(self.target.all_revision_ids())
2754
result_set = set(source_ids).difference(target_ids)
2755
return self.source.revision_ids_to_search_result(result_set)
2758
class InterModel1and2(InterRepository):
2761
def _get_repo_format_to_test(self):
2765
def is_compatible(source, target):
2766
if not source.supports_rich_root() and target.supports_rich_root():
2772
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2773
"""See InterRepository.fetch()."""
2774
from bzrlib.fetch import Model1toKnit2Fetcher
2775
f = Model1toKnit2Fetcher(to_repository=self.target,
2776
from_repository=self.source,
2777
last_revision=revision_id,
2778
pb=pb, find_ghosts=find_ghosts)
2779
return f.count_copied, f.failed_revisions
2782
def copy_content(self, revision_id=None):
2783
"""Make a complete copy of the content in self into destination.
2785
This is a destructive operation! Do not use it on existing
2788
:param revision_id: Only copy the content needed to construct
2789
revision_id and its parents.
2792
self.target.set_make_working_trees(self.source.make_working_trees())
2793
except NotImplementedError:
2795
# but don't bother fetching if we have the needed data now.
2796
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2797
self.target.has_revision(revision_id)):
2799
self.target.fetch(self.source, revision_id=revision_id)
2802
class InterKnit1and2(InterKnitRepo):
2805
def _get_repo_format_to_test(self):
2809
def is_compatible(source, target):
2810
"""Be compatible with Knit1 source and Knit3 target"""
2811
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2813
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2814
RepositoryFormatKnit3)
2815
from bzrlib.repofmt.pack_repo import (
2816
RepositoryFormatKnitPack1,
2817
RepositoryFormatKnitPack3,
2818
RepositoryFormatPackDevelopment0,
2819
RepositoryFormatPackDevelopment0Subtree,
2822
RepositoryFormatKnit1,
2823
RepositoryFormatKnitPack1,
2824
RepositoryFormatPackDevelopment0,
2827
RepositoryFormatKnit3,
2828
RepositoryFormatKnitPack3,
2829
RepositoryFormatPackDevelopment0Subtree,
2831
return (isinstance(source._format, nosubtrees) and
2832
isinstance(target._format, subtrees))
2833
except AttributeError:
2837
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2838
"""See InterRepository.fetch()."""
2839
from bzrlib.fetch import Knit1to2Fetcher
2840
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2841
self.source, self.source._format, self.target,
2842
self.target._format)
2843
f = Knit1to2Fetcher(to_repository=self.target,
2844
from_repository=self.source,
2845
last_revision=revision_id,
2846
pb=pb, find_ghosts=find_ghosts)
2847
return f.count_copied, f.failed_revisions
2850
class InterDifferingSerializer(InterKnitRepo):
2853
def _get_repo_format_to_test(self):
2857
def is_compatible(source, target):
2858
"""Be compatible with Knit2 source and Knit3 target"""
2859
if source.supports_rich_root() != target.supports_rich_root():
2861
# Ideally, we'd support fetching if the source had no tree references
2862
# even if it supported them...
2863
if (getattr(source, '_format.supports_tree_reference', False) and
2864
not getattr(target, '_format.supports_tree_reference', False)):
2869
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2870
"""See InterRepository.fetch()."""
2871
revision_ids = self.target.search_missing_revision_ids(self.source,
2872
revision_id, find_ghosts=find_ghosts).get_keys()
2873
revision_ids = tsort.topo_sort(
2874
self.source.get_graph().get_parent_map(revision_ids))
2875
def revisions_iterator():
2876
for current_revision_id in revision_ids:
2877
revision = self.source.get_revision(current_revision_id)
2878
tree = self.source.revision_tree(current_revision_id)
2880
signature = self.source.get_signature_text(
2881
current_revision_id)
2882
except errors.NoSuchRevision:
2884
yield revision, tree, signature
2886
my_pb = ui.ui_factory.nested_progress_bar()
2891
install_revisions(self.target, revisions_iterator(),
2892
len(revision_ids), pb)
2894
if my_pb is not None:
2896
return len(revision_ids), 0
2899
class InterRemoteToOther(InterRepository):
2901
def __init__(self, source, target):
2902
InterRepository.__init__(self, source, target)
2903
self._real_inter = None
2906
def is_compatible(source, target):
2907
if not isinstance(source, remote.RemoteRepository):
2909
# Is source's model compatible with target's model?
2910
source._ensure_real()
2911
real_source = source._real_repository
2912
if isinstance(real_source, remote.RemoteRepository):
2913
raise NotImplementedError(
2914
"We don't support remote repos backed by remote repos yet.")
2915
return InterRepository._same_model(real_source, target)
2918
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2919
"""See InterRepository.fetch()."""
2920
from bzrlib.fetch import RemoteToOtherFetcher
2921
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2922
self.source, self.target, self.target._format)
2923
# TODO: jam 20070210 This should be an assert, not a translate
2924
revision_id = osutils.safe_revision_id(revision_id)
2925
f = RemoteToOtherFetcher(to_repository=self.target,
2926
from_repository=self.source,
2927
last_revision=revision_id,
2928
pb=pb, find_ghosts=find_ghosts)
2929
return f.count_copied, f.failed_revisions
2932
def _get_repo_format_to_test(self):
2936
class InterOtherToRemote(InterRepository):
2938
def __init__(self, source, target):
2939
InterRepository.__init__(self, source, target)
2940
self._real_inter = None
2943
def is_compatible(source, target):
2944
if isinstance(target, remote.RemoteRepository):
2948
def _ensure_real_inter(self):
2949
if self._real_inter is None:
2950
self.target._ensure_real()
2951
real_target = self.target._real_repository
2952
self._real_inter = InterRepository.get(self.source, real_target)
2954
def copy_content(self, revision_id=None):
2955
self._ensure_real_inter()
2956
self._real_inter.copy_content(revision_id=revision_id)
2958
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2959
self._ensure_real_inter()
2960
self._real_inter.fetch(revision_id=revision_id, pb=pb,
2961
find_ghosts=find_ghosts)
2964
def _get_repo_format_to_test(self):
2968
InterRepository.register_optimiser(InterDifferingSerializer)
2969
InterRepository.register_optimiser(InterSameDataRepository)
2970
InterRepository.register_optimiser(InterWeaveRepo)
2971
InterRepository.register_optimiser(InterKnitRepo)
2972
InterRepository.register_optimiser(InterModel1and2)
2973
InterRepository.register_optimiser(InterKnit1and2)
2974
InterRepository.register_optimiser(InterPackRepo)
2975
InterRepository.register_optimiser(InterRemoteToOther)
2976
InterRepository.register_optimiser(InterOtherToRemote)
2979
class CopyConverter(object):
2980
"""A repository conversion tool which just performs a copy of the content.
2982
This is slow but quite reliable.
2985
def __init__(self, target_format):
2986
"""Create a CopyConverter.
2988
:param target_format: The format the resulting repository should be.
2990
self.target_format = target_format
2992
def convert(self, repo, pb):
2993
"""Perform the conversion of to_convert, giving feedback via pb.
2995
:param to_convert: The disk object to convert.
2996
:param pb: a progress bar to use for progress information.
3001
# this is only useful with metadir layouts - separated repo content.
3002
# trigger an assertion if not such
3003
repo._format.get_format_string()
3004
self.repo_dir = repo.bzrdir
3005
self.step('Moving repository to repository.backup')
3006
self.repo_dir.transport.move('repository', 'repository.backup')
3007
backup_transport = self.repo_dir.transport.clone('repository.backup')
3008
repo._format.check_conversion_target(self.target_format)
3009
self.source_repo = repo._format.open(self.repo_dir,
3011
_override_transport=backup_transport)
3012
self.step('Creating new repository')
3013
converted = self.target_format.initialize(self.repo_dir,
3014
self.source_repo.is_shared())
3015
converted.lock_write()
3017
self.step('Copying content into repository.')
3018
self.source_repo.copy_content_into(converted)
3021
self.step('Deleting old repository content.')
3022
self.repo_dir.transport.delete_tree('repository.backup')
3023
self.pb.note('repository converted')
3025
def step(self, message):
3026
"""Update the pb by a step."""
3028
self.pb.update(message, self.count, self.total)
3040
def _unescaper(match, _map=_unescape_map):
3041
code = match.group(1)
3045
if not code.startswith('#'):
3047
return unichr(int(code[1:])).encode('utf8')
3053
def _unescape_xml(data):
3054
"""Unescape predefined XML entities in a string of data."""
3056
if _unescape_re is None:
3057
_unescape_re = re.compile('\&([^;]*);')
3058
return _unescape_re.sub(_unescaper, data)
3061
class _VersionedFileChecker(object):
3063
def __init__(self, repository):
3064
self.repository = repository
3065
self.text_index = self.repository._generate_text_key_index()
3067
def calculate_file_version_parents(self, revision_id, file_id):
3068
"""Calculate the correct parents for a file version according to
3071
parent_keys = self.text_index[(file_id, revision_id)]
3072
if parent_keys == [_mod_revision.NULL_REVISION]:
3074
# strip the file_id, for the weave api
3075
return tuple([revision_id for file_id, revision_id in parent_keys])
3077
def check_file_version_parents(self, weave, file_id):
3078
"""Check the parents stored in a versioned file are correct.
3080
It also detects file versions that are not referenced by their
3081
corresponding revision's inventory.
3083
:returns: A tuple of (wrong_parents, dangling_file_versions).
3084
wrong_parents is a dict mapping {revision_id: (stored_parents,
3085
correct_parents)} for each revision_id where the stored parents
3086
are not correct. dangling_file_versions is a set of (file_id,
3087
revision_id) tuples for versions that are present in this versioned
3088
file, but not used by the corresponding inventory.
3091
unused_versions = set()
3092
versions = weave.versions()
3093
parent_map = weave.get_parent_map(versions)
3094
for num, revision_id in enumerate(versions):
3096
correct_parents = self.calculate_file_version_parents(
3097
revision_id, file_id)
3099
# The version is not part of the used keys.
3100
unused_versions.add(revision_id)
3103
knit_parents = tuple(parent_map[revision_id])
3104
except errors.RevisionNotPresent:
3106
if correct_parents != knit_parents:
3107
wrong_parents[revision_id] = (knit_parents, correct_parents)
3108
return wrong_parents, unused_versions
3111
def _old_get_graph(repository, revision_id):
3112
"""DO NOT USE. That is all. I'm serious."""
3113
graph = repository.get_graph()
3114
revision_graph = dict(((key, value) for key, value in
3115
graph.iter_ancestry([revision_id]) if value is not None))
3116
return _strip_NULL_ghosts(revision_graph)
3119
def _strip_NULL_ghosts(revision_graph):
3120
"""Also don't use this. more compatibility code for unmigrated clients."""
3121
# Filter ghosts, and null:
3122
if _mod_revision.NULL_REVISION in revision_graph:
3123
del revision_graph[_mod_revision.NULL_REVISION]
3124
for key, parents in revision_graph.items():
3125
revision_graph[key] = tuple(parent for parent in parents if parent
3127
return revision_graph
added
removed
bzrlib/repository.py
