1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
39
revision as _mod_revision,
45
from bzrlib.bundle import serializer
46
from bzrlib.revisiontree import RevisionTree
47
from bzrlib.store.versioned import VersionedFileStore
48
from bzrlib.store.text import TextStore
49
from bzrlib.testament import Testament
50
from bzrlib.util import bencode
53
from bzrlib.decorators import needs_read_lock, needs_write_lock
54
from bzrlib.inter import InterObject
55
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
56
from bzrlib.symbol_versioning import (
59
from bzrlib.trace import mutter, mutter_callsite, note, warning
62
# Old formats display a warning, but only once
63
_deprecation_warning_done = False
66
class CommitBuilder(object):
67
"""Provides an interface to build up a commit.
69
This allows describing a tree to be committed without needing to
70
know the internals of the format of the repository.
73
# all clients should supply tree roots.
74
record_root_entry = True
75
# the default CommitBuilder does not manage trees whose root is versioned.
76
_versioned_root = False
78
def __init__(self, repository, parents, config, timestamp=None,
79
timezone=None, committer=None, revprops=None,
81
"""Initiate a CommitBuilder.
83
:param repository: Repository to commit to.
84
:param parents: Revision ids of the parents of the new revision.
85
:param config: Configuration to use.
86
:param timestamp: Optional timestamp recorded for commit.
87
:param timezone: Optional timezone for timestamp.
88
:param committer: Optional committer to set for commit.
89
:param revprops: Optional dictionary of revision properties.
90
:param revision_id: Optional revision id.
95
self._committer = self._config.username()
97
self._committer = committer
99
self.new_inventory = Inventory(None)
100
self._new_revision_id = revision_id
101
self.parents = parents
102
self.repository = repository
105
if revprops is not None:
106
self._revprops.update(revprops)
108
if timestamp is None:
109
timestamp = time.time()
110
# Restrict resolution to 1ms
111
self._timestamp = round(timestamp, 3)
114
self._timezone = osutils.local_time_offset()
116
self._timezone = int(timezone)
118
self._generate_revision_if_needed()
119
self.__heads = graph.HeadsCache(repository.get_graph()).heads
121
def commit(self, message):
122
"""Make the actual commit.
124
:return: The revision id of the recorded revision.
126
rev = _mod_revision.Revision(
127
timestamp=self._timestamp,
128
timezone=self._timezone,
129
committer=self._committer,
131
inventory_sha1=self.inv_sha1,
132
revision_id=self._new_revision_id,
133
properties=self._revprops)
134
rev.parent_ids = self.parents
135
self.repository.add_revision(self._new_revision_id, rev,
136
self.new_inventory, self._config)
137
self.repository.commit_write_group()
138
return self._new_revision_id
141
"""Abort the commit that is being built.
143
self.repository.abort_write_group()
145
def revision_tree(self):
146
"""Return the tree that was just committed.
148
After calling commit() this can be called to get a RevisionTree
149
representing the newly committed tree. This is preferred to
150
calling Repository.revision_tree() because that may require
151
deserializing the inventory, while we already have a copy in
154
return RevisionTree(self.repository, self.new_inventory,
155
self._new_revision_id)
157
def finish_inventory(self):
158
"""Tell the builder that the inventory is finished."""
159
if self.new_inventory.root is None:
160
raise AssertionError('Root entry should be supplied to'
161
' record_entry_contents, as of bzr 0.10.',
162
DeprecationWarning, stacklevel=2)
163
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
164
self.new_inventory.revision_id = self._new_revision_id
165
self.inv_sha1 = self.repository.add_inventory(
166
self._new_revision_id,
171
def _gen_revision_id(self):
172
"""Return new revision-id."""
173
return generate_ids.gen_revision_id(self._config.username(),
176
def _generate_revision_if_needed(self):
177
"""Create a revision id if None was supplied.
179
If the repository can not support user-specified revision ids
180
they should override this function and raise CannotSetRevisionId
181
if _new_revision_id is not None.
183
:raises: CannotSetRevisionId
185
if self._new_revision_id is None:
186
self._new_revision_id = self._gen_revision_id()
187
self.random_revid = True
189
self.random_revid = False
191
def _heads(self, file_id, revision_ids):
192
"""Calculate the graph heads for revision_ids in the graph of file_id.
194
This can use either a per-file graph or a global revision graph as we
195
have an identity relationship between the two graphs.
197
return self.__heads(revision_ids)
199
def _check_root(self, ie, parent_invs, tree):
200
"""Helper for record_entry_contents.
202
:param ie: An entry being added.
203
:param parent_invs: The inventories of the parent revisions of the
205
:param tree: The tree that is being committed.
207
# In this revision format, root entries have no knit or weave When
208
# serializing out to disk and back in root.revision is always
210
ie.revision = self._new_revision_id
212
def _get_delta(self, ie, basis_inv, path):
213
"""Get a delta against the basis inventory for ie."""
214
if ie.file_id not in basis_inv:
216
return (None, path, ie.file_id, ie)
217
elif ie != basis_inv[ie.file_id]:
219
# TODO: avoid tis id2path call.
220
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
225
def record_entry_contents(self, ie, parent_invs, path, tree,
227
"""Record the content of ie from tree into the commit if needed.
229
Side effect: sets ie.revision when unchanged
231
:param ie: An inventory entry present in the commit.
232
:param parent_invs: The inventories of the parent revisions of the
234
:param path: The path the entry is at in the tree.
235
:param tree: The tree which contains this entry and should be used to
237
:param content_summary: Summary data from the tree about the paths
238
content - stat, length, exec, sha/link target. This is only
239
accessed when the entry has a revision of None - that is when it is
240
a candidate to commit.
241
:return: A tuple (change_delta, version_recorded). change_delta is
242
an inventory_delta change for this entry against the basis tree of
243
the commit, or None if no change occured against the basis tree.
244
version_recorded is True if a new version of the entry has been
245
recorded. For instance, committing a merge where a file was only
246
changed on the other side will return (delta, False).
248
if self.new_inventory.root is None:
249
if ie.parent_id is not None:
250
raise errors.RootMissing()
251
self._check_root(ie, parent_invs, tree)
252
if ie.revision is None:
253
kind = content_summary[0]
255
# ie is carried over from a prior commit
257
# XXX: repository specific check for nested tree support goes here - if
258
# the repo doesn't want nested trees we skip it ?
259
if (kind == 'tree-reference' and
260
not self.repository._format.supports_tree_reference):
261
# mismatch between commit builder logic and repository:
262
# this needs the entry creation pushed down into the builder.
263
raise NotImplementedError('Missing repository subtree support.')
264
self.new_inventory.add(ie)
266
# TODO: slow, take it out of the inner loop.
268
basis_inv = parent_invs[0]
270
basis_inv = Inventory(root_id=None)
272
# ie.revision is always None if the InventoryEntry is considered
273
# for committing. We may record the previous parents revision if the
274
# content is actually unchanged against a sole head.
275
if ie.revision is not None:
276
if not self._versioned_root and path == '':
277
# repositories that do not version the root set the root's
278
# revision to the new commit even when no change occurs, and
279
# this masks when a change may have occurred against the basis,
280
# so calculate if one happened.
281
if ie.file_id in basis_inv:
282
delta = (basis_inv.id2path(ie.file_id), path,
286
delta = (None, path, ie.file_id, ie)
289
# we don't need to commit this, because the caller already
290
# determined that an existing revision of this file is
292
return None, (ie.revision == self._new_revision_id)
293
# XXX: Friction: parent_candidates should return a list not a dict
294
# so that we don't have to walk the inventories again.
295
parent_candiate_entries = ie.parent_candidates(parent_invs)
296
head_set = self._heads(ie.file_id, parent_candiate_entries.keys())
298
for inv in parent_invs:
299
if ie.file_id in inv:
300
old_rev = inv[ie.file_id].revision
301
if old_rev in head_set:
302
heads.append(inv[ie.file_id].revision)
303
head_set.remove(inv[ie.file_id].revision)
306
# now we check to see if we need to write a new record to the
308
# We write a new entry unless there is one head to the ancestors, and
309
# the kind-derived content is unchanged.
311
# Cheapest check first: no ancestors, or more the one head in the
312
# ancestors, we write a new node.
316
# There is a single head, look it up for comparison
317
parent_entry = parent_candiate_entries[heads[0]]
318
# if the non-content specific data has changed, we'll be writing a
320
if (parent_entry.parent_id != ie.parent_id or
321
parent_entry.name != ie.name):
323
# now we need to do content specific checks:
325
# if the kind changed the content obviously has
326
if kind != parent_entry.kind:
329
if content_summary[2] is None:
330
raise ValueError("Files must not have executable = None")
332
if (# if the file length changed we have to store:
333
parent_entry.text_size != content_summary[1] or
334
# if the exec bit has changed we have to store:
335
parent_entry.executable != content_summary[2]):
337
elif parent_entry.text_sha1 == content_summary[3]:
338
# all meta and content is unchanged (using a hash cache
339
# hit to check the sha)
340
ie.revision = parent_entry.revision
341
ie.text_size = parent_entry.text_size
342
ie.text_sha1 = parent_entry.text_sha1
343
ie.executable = parent_entry.executable
344
return self._get_delta(ie, basis_inv, path), False
346
# Either there is only a hash change(no hash cache entry,
347
# or same size content change), or there is no change on
349
# Provide the parent's hash to the store layer, so that the
350
# content is unchanged we will not store a new node.
351
nostore_sha = parent_entry.text_sha1
353
# We want to record a new node regardless of the presence or
354
# absence of a content change in the file.
356
ie.executable = content_summary[2]
357
lines = tree.get_file(ie.file_id, path).readlines()
359
ie.text_sha1, ie.text_size = self._add_text_to_weave(
360
ie.file_id, lines, heads, nostore_sha)
361
except errors.ExistingContent:
362
# Turns out that the file content was unchanged, and we were
363
# only going to store a new node if it was changed. Carry over
365
ie.revision = parent_entry.revision
366
ie.text_size = parent_entry.text_size
367
ie.text_sha1 = parent_entry.text_sha1
368
ie.executable = parent_entry.executable
369
return self._get_delta(ie, basis_inv, path), False
370
elif kind == 'directory':
372
# all data is meta here, nothing specific to directory, so
374
ie.revision = parent_entry.revision
375
return self._get_delta(ie, basis_inv, path), False
377
self._add_text_to_weave(ie.file_id, lines, heads, None)
378
elif kind == 'symlink':
379
current_link_target = content_summary[3]
381
# symlink target is not generic metadata, check if it has
383
if current_link_target != parent_entry.symlink_target:
386
# unchanged, carry over.
387
ie.revision = parent_entry.revision
388
ie.symlink_target = parent_entry.symlink_target
389
return self._get_delta(ie, basis_inv, path), False
390
ie.symlink_target = current_link_target
392
self._add_text_to_weave(ie.file_id, lines, heads, None)
393
elif kind == 'tree-reference':
395
if content_summary[3] != parent_entry.reference_revision:
398
# unchanged, carry over.
399
ie.reference_revision = parent_entry.reference_revision
400
ie.revision = parent_entry.revision
401
return self._get_delta(ie, basis_inv, path), False
402
ie.reference_revision = content_summary[3]
404
self._add_text_to_weave(ie.file_id, lines, heads, None)
406
raise NotImplementedError('unknown kind')
407
ie.revision = self._new_revision_id
408
return self._get_delta(ie, basis_inv, path), True
410
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
411
versionedfile = self.repository.weave_store.get_weave_or_empty(
412
file_id, self.repository.get_transaction())
413
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
414
# than add_lines, and allows committing when a parent is ghosted for
416
# Note: as we read the content directly from the tree, we know its not
417
# been turned into unicode or badly split - but a broken tree
418
# implementation could give us bad output from readlines() so this is
419
# not a guarantee of safety. What would be better is always checking
420
# the content during test suite execution. RBC 20070912
421
return versionedfile.add_lines_with_ghosts(
422
self._new_revision_id, parents, new_lines,
423
nostore_sha=nostore_sha, random_id=self.random_revid,
424
check_content=False)[0:2]
427
class RootCommitBuilder(CommitBuilder):
428
"""This commitbuilder actually records the root id"""
430
# the root entry gets versioned properly by this builder.
431
_versioned_root = True
433
def _check_root(self, ie, parent_invs, tree):
434
"""Helper for record_entry_contents.
436
:param ie: An entry being added.
437
:param parent_invs: The inventories of the parent revisions of the
439
:param tree: The tree that is being committed.
443
######################################################################
446
class Repository(object):
447
"""Repository holding history for one or more branches.
449
The repository holds and retrieves historical information including
450
revisions and file history. It's normally accessed only by the Branch,
451
which views a particular line of development through that history.
453
The Repository builds on top of Stores and a Transport, which respectively
454
describe the disk data format and the way of accessing the (possibly
457
:ivar _transport: Transport for file access to repository, typically
458
pointing to .bzr/repository.
461
# What class to use for a CommitBuilder. Often its simpler to change this
462
# in a Repository class subclass rather than to override
463
# get_commit_builder.
464
_commit_builder_class = CommitBuilder
465
# The search regex used by xml based repositories to determine what things
466
# where changed in a single commit.
467
_file_ids_altered_regex = lazy_regex.lazy_compile(
468
r'file_id="(?P<file_id>[^"]+)"'
469
r'.* revision="(?P<revision_id>[^"]+)"'
472
def abort_write_group(self):
473
"""Commit the contents accrued within the current write group.
475
:seealso: start_write_group.
477
if self._write_group is not self.get_transaction():
478
# has an unlock or relock occured ?
479
raise errors.BzrError('mismatched lock context and write group.')
480
self._abort_write_group()
481
self._write_group = None
483
def _abort_write_group(self):
484
"""Template method for per-repository write group cleanup.
486
This is called during abort before the write group is considered to be
487
finished and should cleanup any internal state accrued during the write
488
group. There is no requirement that data handed to the repository be
489
*not* made available - this is not a rollback - but neither should any
490
attempt be made to ensure that data added is fully commited. Abort is
491
invoked when an error has occured so futher disk or network operations
492
may not be possible or may error and if possible should not be
496
def add_inventory(self, revision_id, inv, parents):
497
"""Add the inventory inv to the repository as revision_id.
499
:param parents: The revision ids of the parents that revision_id
500
is known to have and are in the repository already.
502
:returns: The validator(which is a sha1 digest, though what is sha'd is
503
repository format specific) of the serialized inventory.
505
if not self.is_in_write_group():
506
raise AssertionError("%r not in write group" % (self,))
507
_mod_revision.check_not_reserved_id(revision_id)
508
if not (inv.revision_id is None or inv.revision_id == revision_id):
509
raise AssertionError(
510
"Mismatch between inventory revision"
511
" id and insertion revid (%r, %r)"
512
% (inv.revision_id, revision_id))
514
raise AssertionError()
515
inv_lines = self._serialise_inventory_to_lines(inv)
516
inv_vf = self.get_inventory_weave()
517
return self._inventory_add_lines(inv_vf, revision_id, parents,
518
inv_lines, check_content=False)
520
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
522
"""Store lines in inv_vf and return the sha1 of the inventory."""
524
for parent in parents:
526
final_parents.append(parent)
527
return inv_vf.add_lines(revision_id, final_parents, lines,
528
check_content=check_content)[0]
530
def add_revision(self, revision_id, rev, inv=None, config=None):
531
"""Add rev to the revision store as revision_id.
533
:param revision_id: the revision id to use.
534
:param rev: The revision object.
535
:param inv: The inventory for the revision. if None, it will be looked
536
up in the inventory storer
537
:param config: If None no digital signature will be created.
538
If supplied its signature_needed method will be used
539
to determine if a signature should be made.
541
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
543
_mod_revision.check_not_reserved_id(revision_id)
544
if config is not None and config.signature_needed():
546
inv = self.get_inventory(revision_id)
547
plaintext = Testament(rev, inv).as_short_text()
548
self.store_revision_signature(
549
gpg.GPGStrategy(config), plaintext, revision_id)
550
inventory_vf = self.get_inventory_weave()
551
if not revision_id in inventory_vf:
553
raise errors.WeaveRevisionNotPresent(revision_id,
556
# yes, this is not suitable for adding with ghosts.
557
rev.inventory_sha1 = self.add_inventory(revision_id, inv,
560
rev.inventory_sha1 = inventory_vf.get_sha1s([revision_id])[0]
561
self._revision_store.add_revision(rev, self.get_transaction())
563
def _add_revision_text(self, revision_id, text):
564
revision = self._revision_store._serializer.read_revision_from_string(
566
self._revision_store._add_revision(revision, StringIO(text),
567
self.get_transaction())
569
def all_revision_ids(self):
570
"""Returns a list of all the revision ids in the repository.
572
This is deprecated because code should generally work on the graph
573
reachable from a particular revision, and ignore any other revisions
574
that might be present. There is no direct replacement method.
576
if 'evil' in debug.debug_flags:
577
mutter_callsite(2, "all_revision_ids is linear with history.")
578
return self._all_revision_ids()
580
def _all_revision_ids(self):
581
"""Returns a list of all the revision ids in the repository.
583
These are in as much topological order as the underlying store can
586
raise NotImplementedError(self._all_revision_ids)
588
def break_lock(self):
589
"""Break a lock if one is present from another instance.
591
Uses the ui factory to ask for confirmation if the lock may be from
594
self.control_files.break_lock()
597
def _eliminate_revisions_not_present(self, revision_ids):
598
"""Check every revision id in revision_ids to see if we have it.
600
Returns a set of the present revisions.
603
graph = self.get_graph()
604
parent_map = graph.get_parent_map(revision_ids)
605
# The old API returned a list, should this actually be a set?
606
return parent_map.keys()
609
def create(a_bzrdir):
610
"""Construct the current default format repository in a_bzrdir."""
611
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
613
def __init__(self, _format, a_bzrdir, control_files,
614
_revision_store, control_store, text_store):
615
"""instantiate a Repository.
617
:param _format: The format of the repository on disk.
618
:param a_bzrdir: The BzrDir of the repository.
620
In the future we will have a single api for all stores for
621
getting file texts, inventories and revisions, then
622
this construct will accept instances of those things.
624
super(Repository, self).__init__()
625
self._format = _format
626
# the following are part of the public API for Repository:
627
self.bzrdir = a_bzrdir
628
self.control_files = control_files
629
self._transport = control_files._transport
630
self.base = self._transport.base
631
self._revision_store = _revision_store
632
# backwards compatibility
633
self.weave_store = text_store
635
self._reconcile_does_inventory_gc = True
636
self._reconcile_fixes_text_parents = False
637
self._reconcile_backsup_inventory = True
638
# not right yet - should be more semantically clear ?
640
self.control_store = control_store
641
self.control_weaves = control_store
642
# TODO: make sure to construct the right store classes, etc, depending
643
# on whether escaping is required.
644
self._warn_if_deprecated()
645
self._write_group = None
648
return '%s(%r)' % (self.__class__.__name__,
651
def has_same_location(self, other):
652
"""Returns a boolean indicating if this repository is at the same
653
location as another repository.
655
This might return False even when two repository objects are accessing
656
the same physical repository via different URLs.
658
if self.__class__ is not other.__class__:
660
return (self._transport.base == other._transport.base)
662
def is_in_write_group(self):
663
"""Return True if there is an open write group.
665
:seealso: start_write_group.
667
return self._write_group is not None
670
return self.control_files.is_locked()
672
def is_write_locked(self):
673
"""Return True if this object is write locked."""
674
return self.is_locked() and self.control_files._lock_mode == 'w'
676
def lock_write(self, token=None):
677
"""Lock this repository for writing.
679
This causes caching within the repository obejct to start accumlating
680
data during reads, and allows a 'write_group' to be obtained. Write
681
groups must be used for actual data insertion.
683
:param token: if this is already locked, then lock_write will fail
684
unless the token matches the existing lock.
685
:returns: a token if this instance supports tokens, otherwise None.
686
:raises TokenLockingNotSupported: when a token is given but this
687
instance doesn't support using token locks.
688
:raises MismatchedToken: if the specified token doesn't match the token
689
of the existing lock.
690
:seealso: start_write_group.
692
A token should be passed in if you know that you have locked the object
693
some other way, and need to synchronise this object's state with that
696
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
698
result = self.control_files.lock_write(token=token)
703
self.control_files.lock_read()
706
def get_physical_lock_status(self):
707
return self.control_files.get_physical_lock_status()
709
def leave_lock_in_place(self):
710
"""Tell this repository not to release the physical lock when this
713
If lock_write doesn't return a token, then this method is not supported.
715
self.control_files.leave_in_place()
717
def dont_leave_lock_in_place(self):
718
"""Tell this repository to release the physical lock when this
719
object is unlocked, even if it didn't originally acquire it.
721
If lock_write doesn't return a token, then this method is not supported.
723
self.control_files.dont_leave_in_place()
726
def gather_stats(self, revid=None, committers=None):
727
"""Gather statistics from a revision id.
729
:param revid: The revision id to gather statistics from, if None, then
730
no revision specific statistics are gathered.
731
:param committers: Optional parameter controlling whether to grab
732
a count of committers from the revision specific statistics.
733
:return: A dictionary of statistics. Currently this contains:
734
committers: The number of committers if requested.
735
firstrev: A tuple with timestamp, timezone for the penultimate left
736
most ancestor of revid, if revid is not the NULL_REVISION.
737
latestrev: A tuple with timestamp, timezone for revid, if revid is
738
not the NULL_REVISION.
739
revisions: The total revision count in the repository.
740
size: An estimate disk size of the repository in bytes.
743
if revid and committers:
744
result['committers'] = 0
745
if revid and revid != _mod_revision.NULL_REVISION:
747
all_committers = set()
748
revisions = self.get_ancestry(revid)
749
# pop the leading None
751
first_revision = None
753
# ignore the revisions in the middle - just grab first and last
754
revisions = revisions[0], revisions[-1]
755
for revision in self.get_revisions(revisions):
756
if not first_revision:
757
first_revision = revision
759
all_committers.add(revision.committer)
760
last_revision = revision
762
result['committers'] = len(all_committers)
763
result['firstrev'] = (first_revision.timestamp,
764
first_revision.timezone)
765
result['latestrev'] = (last_revision.timestamp,
766
last_revision.timezone)
768
# now gather global repository information
769
if self.bzrdir.root_transport.listable():
770
c, t = self._revision_store.total_size(self.get_transaction())
771
result['revisions'] = c
775
def find_branches(self, using=False):
776
"""Find branches underneath this repository.
778
This will include branches inside other branches.
780
:param using: If True, list only branches using this repository.
782
if using and not self.is_shared():
784
return [self.bzrdir.open_branch()]
785
except errors.NotBranchError:
787
class Evaluator(object):
790
self.first_call = True
792
def __call__(self, bzrdir):
793
# On the first call, the parameter is always the bzrdir
794
# containing the current repo.
795
if not self.first_call:
797
repository = bzrdir.open_repository()
798
except errors.NoRepositoryPresent:
801
return False, (None, repository)
802
self.first_call = False
804
value = (bzrdir.open_branch(), None)
805
except errors.NotBranchError:
810
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
811
self.bzrdir.root_transport, evaluate=Evaluator()):
812
if branch is not None:
813
branches.append(branch)
814
if not using and repository is not None:
815
branches.extend(repository.find_branches())
818
def get_data_stream(self, revision_ids):
819
raise NotImplementedError(self.get_data_stream)
821
def get_data_stream_for_search(self, search_result):
822
"""Get a data stream that can be inserted to a repository.
824
:param search_result: A bzrlib.graph.SearchResult selecting the
826
:return: A data stream that can be inserted into a repository using
829
raise NotImplementedError(self.get_data_stream_for_search)
831
def insert_data_stream(self, stream):
832
"""XXX What does this really do?
834
Is it a substitute for fetch?
835
Should it manage its own write group ?
837
for item_key, bytes in stream:
838
if item_key[0] == 'file':
839
(file_id,) = item_key[1:]
840
knit = self.weave_store.get_weave_or_empty(
841
file_id, self.get_transaction())
842
elif item_key == ('inventory',):
843
knit = self.get_inventory_weave()
844
elif item_key == ('revisions',):
845
knit = self._revision_store.get_revision_file(
846
self.get_transaction())
847
elif item_key == ('signatures',):
848
knit = self._revision_store.get_signature_file(
849
self.get_transaction())
851
raise errors.RepositoryDataStreamError(
852
"Unrecognised data stream key '%s'" % (item_key,))
853
decoded_list = bencode.bdecode(bytes)
854
format = decoded_list.pop(0)
857
for version, options, parents, some_bytes in decoded_list:
858
data_list.append((version, options, len(some_bytes), parents))
859
knit_bytes += some_bytes
860
buffer = StringIO(knit_bytes)
861
def reader_func(count):
865
return buffer.read(count)
866
knit.insert_data_stream(
867
(format, data_list, reader_func))
870
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
871
"""Return the revision ids that other has that this does not.
873
These are returned in topological order.
875
revision_id: only return revision ids included by revision_id.
877
return InterRepository.get(other, self).search_missing_revision_ids(
878
revision_id, find_ghosts)
880
@deprecated_method(symbol_versioning.one_two)
882
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
883
"""Return the revision ids that other has that this does not.
885
These are returned in topological order.
887
revision_id: only return revision ids included by revision_id.
889
keys = self.search_missing_revision_ids(
890
other, revision_id, find_ghosts).get_keys()
893
parents = other.get_graph().get_parent_map(keys)
896
return tsort.topo_sort(parents)
900
"""Open the repository rooted at base.
902
For instance, if the repository is at URL/.bzr/repository,
903
Repository.open(URL) -> a Repository instance.
905
control = bzrdir.BzrDir.open(base)
906
return control.open_repository()
908
def copy_content_into(self, destination, revision_id=None):
909
"""Make a complete copy of the content in self into destination.
911
This is a destructive operation! Do not use it on existing
914
return InterRepository.get(self, destination).copy_content(revision_id)
916
def commit_write_group(self):
917
"""Commit the contents accrued within the current write group.
919
:seealso: start_write_group.
921
if self._write_group is not self.get_transaction():
922
# has an unlock or relock occured ?
923
raise errors.BzrError('mismatched lock context %r and '
925
(self.get_transaction(), self._write_group))
926
self._commit_write_group()
927
self._write_group = None
929
def _commit_write_group(self):
930
"""Template method for per-repository write group cleanup.
932
This is called before the write group is considered to be
933
finished and should ensure that all data handed to the repository
934
for writing during the write group is safely committed (to the
935
extent possible considering file system caching etc).
938
def fetch(self, source, revision_id=None, pb=None, find_ghosts=False):
939
"""Fetch the content required to construct revision_id from source.
941
If revision_id is None all content is copied.
942
:param find_ghosts: Find and copy revisions in the source that are
943
ghosts in the target (and not reachable directly by walking out to
944
the first-present revision in target from revision_id).
946
# fast path same-url fetch operations
947
if self.has_same_location(source):
948
# check that last_revision is in 'from' and then return a
950
if (revision_id is not None and
951
not _mod_revision.is_null(revision_id)):
952
self.get_revision(revision_id)
954
inter = InterRepository.get(source, self)
956
return inter.fetch(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts)
957
except NotImplementedError:
958
raise errors.IncompatibleRepositories(source, self)
960
def create_bundle(self, target, base, fileobj, format=None):
961
return serializer.write_bundle(self, target, base, fileobj, format)
963
def get_commit_builder(self, branch, parents, config, timestamp=None,
964
timezone=None, committer=None, revprops=None,
966
"""Obtain a CommitBuilder for this repository.
968
:param branch: Branch to commit to.
969
:param parents: Revision ids of the parents of the new revision.
970
:param config: Configuration to use.
971
:param timestamp: Optional timestamp recorded for commit.
972
:param timezone: Optional timezone for timestamp.
973
:param committer: Optional committer to set for commit.
974
:param revprops: Optional dictionary of revision properties.
975
:param revision_id: Optional revision id.
977
result = self._commit_builder_class(self, parents, config,
978
timestamp, timezone, committer, revprops, revision_id)
979
self.start_write_group()
983
if (self.control_files._lock_count == 1 and
984
self.control_files._lock_mode == 'w'):
985
if self._write_group is not None:
986
self.abort_write_group()
987
self.control_files.unlock()
988
raise errors.BzrError(
989
'Must end write groups before releasing write locks.')
990
self.control_files.unlock()
993
def clone(self, a_bzrdir, revision_id=None):
994
"""Clone this repository into a_bzrdir using the current format.
996
Currently no check is made that the format of this repository and
997
the bzrdir format are compatible. FIXME RBC 20060201.
999
:return: The newly created destination repository.
1001
# TODO: deprecate after 0.16; cloning this with all its settings is
1002
# probably not very useful -- mbp 20070423
1003
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
1004
self.copy_content_into(dest_repo, revision_id)
1007
def start_write_group(self):
1008
"""Start a write group in the repository.
1010
Write groups are used by repositories which do not have a 1:1 mapping
1011
between file ids and backend store to manage the insertion of data from
1012
both fetch and commit operations.
1014
A write lock is required around the start_write_group/commit_write_group
1015
for the support of lock-requiring repository formats.
1017
One can only insert data into a repository inside a write group.
1021
if not self.is_write_locked():
1022
raise errors.NotWriteLocked(self)
1023
if self._write_group:
1024
raise errors.BzrError('already in a write group')
1025
self._start_write_group()
1026
# so we can detect unlock/relock - the write group is now entered.
1027
self._write_group = self.get_transaction()
1029
def _start_write_group(self):
1030
"""Template method for per-repository write group startup.
1032
This is called before the write group is considered to be
1037
def sprout(self, to_bzrdir, revision_id=None):
1038
"""Create a descendent repository for new development.
1040
Unlike clone, this does not copy the settings of the repository.
1042
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
1043
dest_repo.fetch(self, revision_id=revision_id)
1046
def _create_sprouting_repo(self, a_bzrdir, shared):
1047
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
1048
# use target default format.
1049
dest_repo = a_bzrdir.create_repository()
1051
# Most control formats need the repository to be specifically
1052
# created, but on some old all-in-one formats it's not needed
1054
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
1055
except errors.UninitializableFormat:
1056
dest_repo = a_bzrdir.open_repository()
1060
def has_revision(self, revision_id):
1061
"""True if this repository has a copy of the revision."""
1062
return revision_id in self.has_revisions((revision_id,))
1064
def has_revisions(self, revision_ids):
1065
"""Probe to find out the presence of multiple revisions.
1067
:param revision_ids: An iterable of revision_ids.
1068
:return: A set of the revision_ids that were present.
1070
raise NotImplementedError(self.has_revisions)
1072
return self._revision_store.has_revision_id(revision_id,
1073
self.get_transaction())
1076
def get_revision(self, revision_id):
1077
"""Return the Revision object for a named revision."""
1078
return self.get_revisions([revision_id])[0]
1081
def get_revision_reconcile(self, revision_id):
1082
"""'reconcile' helper routine that allows access to a revision always.
1084
This variant of get_revision does not cross check the weave graph
1085
against the revision one as get_revision does: but it should only
1086
be used by reconcile, or reconcile-alike commands that are correcting
1087
or testing the revision graph.
1089
return self._get_revisions([revision_id])[0]
1092
def get_revisions(self, revision_ids):
1093
"""Get many revisions at once."""
1094
return self._get_revisions(revision_ids)
1097
def _get_revisions(self, revision_ids):
1098
"""Core work logic to get many revisions without sanity checks."""
1099
for rev_id in revision_ids:
1100
if not rev_id or not isinstance(rev_id, basestring):
1101
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1102
revs = self._revision_store.get_revisions(revision_ids,
1103
self.get_transaction())
1107
def get_revision_xml(self, revision_id):
1108
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1109
# would have already do it.
1110
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1111
rev = self.get_revision(revision_id)
1112
rev_tmp = StringIO()
1113
# the current serializer..
1114
self._revision_store._serializer.write_revision(rev, rev_tmp)
1116
return rev_tmp.getvalue()
1118
def get_deltas_for_revisions(self, revisions):
1119
"""Produce a generator of revision deltas.
1121
Note that the input is a sequence of REVISIONS, not revision_ids.
1122
Trees will be held in memory until the generator exits.
1123
Each delta is relative to the revision's lefthand predecessor.
1125
required_trees = set()
1126
for revision in revisions:
1127
required_trees.add(revision.revision_id)
1128
required_trees.update(revision.parent_ids[:1])
1129
trees = dict((t.get_revision_id(), t) for
1130
t in self.revision_trees(required_trees))
1131
for revision in revisions:
1132
if not revision.parent_ids:
1133
old_tree = self.revision_tree(None)
1135
old_tree = trees[revision.parent_ids[0]]
1136
yield trees[revision.revision_id].changes_from(old_tree)
1139
def get_revision_delta(self, revision_id):
1140
"""Return the delta for one revision.
1142
The delta is relative to the left-hand predecessor of the
1145
r = self.get_revision(revision_id)
1146
return list(self.get_deltas_for_revisions([r]))[0]
1149
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1150
signature = gpg_strategy.sign(plaintext)
1151
self.add_signature_text(revision_id, signature)
1154
def add_signature_text(self, revision_id, signature):
1155
self._revision_store.add_revision_signature_text(revision_id,
1157
self.get_transaction())
1159
def find_text_key_references(self):
1160
"""Find the text key references within the repository.
1162
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1163
revision_ids. Each altered file-ids has the exact revision_ids that
1164
altered it listed explicitly.
1165
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1166
to whether they were referred to by the inventory of the
1167
revision_id that they contain. The inventory texts from all present
1168
revision ids are assessed to generate this report.
1170
revision_ids = self.all_revision_ids()
1171
w = self.get_inventory_weave()
1172
pb = ui.ui_factory.nested_progress_bar()
1174
return self._find_text_key_references_from_xml_inventory_lines(
1175
w.iter_lines_added_or_present_in_versions(revision_ids, pb=pb))
1179
def _find_text_key_references_from_xml_inventory_lines(self,
1181
"""Core routine for extracting references to texts from inventories.
1183
This performs the translation of xml lines to revision ids.
1185
:param line_iterator: An iterator of lines, origin_version_id
1186
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1187
to whether they were referred to by the inventory of the
1188
revision_id that they contain. Note that if that revision_id was
1189
not part of the line_iterator's output then False will be given -
1190
even though it may actually refer to that key.
1192
if not self._serializer.support_altered_by_hack:
1193
raise AssertionError(
1194
"_find_text_key_references_from_xml_inventory_lines only "
1195
"supported for branches which store inventory as unnested xml"
1196
", not on %r" % self)
1199
# this code needs to read every new line in every inventory for the
1200
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1201
# not present in one of those inventories is unnecessary but not
1202
# harmful because we are filtering by the revision id marker in the
1203
# inventory lines : we only select file ids altered in one of those
1204
# revisions. We don't need to see all lines in the inventory because
1205
# only those added in an inventory in rev X can contain a revision=X
1207
unescape_revid_cache = {}
1208
unescape_fileid_cache = {}
1210
# jam 20061218 In a big fetch, this handles hundreds of thousands
1211
# of lines, so it has had a lot of inlining and optimizing done.
1212
# Sorry that it is a little bit messy.
1213
# Move several functions to be local variables, since this is a long
1215
search = self._file_ids_altered_regex.search
1216
unescape = _unescape_xml
1217
setdefault = result.setdefault
1218
for line, version_id in line_iterator:
1219
match = search(line)
1222
# One call to match.group() returning multiple items is quite a
1223
# bit faster than 2 calls to match.group() each returning 1
1224
file_id, revision_id = match.group('file_id', 'revision_id')
1226
# Inlining the cache lookups helps a lot when you make 170,000
1227
# lines and 350k ids, versus 8.4 unique ids.
1228
# Using a cache helps in 2 ways:
1229
# 1) Avoids unnecessary decoding calls
1230
# 2) Re-uses cached strings, which helps in future set and
1232
# (2) is enough that removing encoding entirely along with
1233
# the cache (so we are using plain strings) results in no
1234
# performance improvement.
1236
revision_id = unescape_revid_cache[revision_id]
1238
unescaped = unescape(revision_id)
1239
unescape_revid_cache[revision_id] = unescaped
1240
revision_id = unescaped
1242
# Note that unconditionally unescaping means that we deserialise
1243
# every fileid, which for general 'pull' is not great, but we don't
1244
# really want to have some many fulltexts that this matters anyway.
1247
file_id = unescape_fileid_cache[file_id]
1249
unescaped = unescape(file_id)
1250
unescape_fileid_cache[file_id] = unescaped
1253
key = (file_id, revision_id)
1254
setdefault(key, False)
1255
if revision_id == version_id:
1259
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1261
"""Helper routine for fileids_altered_by_revision_ids.
1263
This performs the translation of xml lines to revision ids.
1265
:param line_iterator: An iterator of lines, origin_version_id
1266
:param revision_ids: The revision ids to filter for. This should be a
1267
set or other type which supports efficient __contains__ lookups, as
1268
the revision id from each parsed line will be looked up in the
1269
revision_ids filter.
1270
:return: a dictionary mapping altered file-ids to an iterable of
1271
revision_ids. Each altered file-ids has the exact revision_ids that
1272
altered it listed explicitly.
1275
setdefault = result.setdefault
1276
for file_id, revision_id in \
1277
self._find_text_key_references_from_xml_inventory_lines(
1278
line_iterator).iterkeys():
1279
# once data is all ensured-consistent; then this is
1280
# if revision_id == version_id
1281
if revision_id in revision_ids:
1282
setdefault(file_id, set()).add(revision_id)
1285
def fileids_altered_by_revision_ids(self, revision_ids, _inv_weave=None):
1286
"""Find the file ids and versions affected by revisions.
1288
:param revisions: an iterable containing revision ids.
1289
:param _inv_weave: The inventory weave from this repository or None.
1290
If None, the inventory weave will be opened automatically.
1291
:return: a dictionary mapping altered file-ids to an iterable of
1292
revision_ids. Each altered file-ids has the exact revision_ids that
1293
altered it listed explicitly.
1295
selected_revision_ids = set(revision_ids)
1296
w = _inv_weave or self.get_inventory_weave()
1297
pb = ui.ui_factory.nested_progress_bar()
1299
return self._find_file_ids_from_xml_inventory_lines(
1300
w.iter_lines_added_or_present_in_versions(
1301
selected_revision_ids, pb=pb),
1302
selected_revision_ids)
1306
def iter_files_bytes(self, desired_files):
1307
"""Iterate through file versions.
1309
Files will not necessarily be returned in the order they occur in
1310
desired_files. No specific order is guaranteed.
1312
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1313
value supplied by the caller as part of desired_files. It should
1314
uniquely identify the file version in the caller's context. (Examples:
1315
an index number or a TreeTransform trans_id.)
1317
bytes_iterator is an iterable of bytestrings for the file. The
1318
kind of iterable and length of the bytestrings are unspecified, but for
1319
this implementation, it is a list of lines produced by
1320
VersionedFile.get_lines().
1322
:param desired_files: a list of (file_id, revision_id, identifier)
1325
transaction = self.get_transaction()
1326
for file_id, revision_id, callable_data in desired_files:
1328
weave = self.weave_store.get_weave(file_id, transaction)
1329
except errors.NoSuchFile:
1330
raise errors.NoSuchIdInRepository(self, file_id)
1331
yield callable_data, weave.get_lines(revision_id)
1333
def _generate_text_key_index(self, text_key_references=None,
1335
"""Generate a new text key index for the repository.
1337
This is an expensive function that will take considerable time to run.
1339
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1340
list of parents, also text keys. When a given key has no parents,
1341
the parents list will be [NULL_REVISION].
1343
# All revisions, to find inventory parents.
1344
if ancestors is None:
1345
graph = self.get_graph()
1346
ancestors = graph.get_parent_map(self.all_revision_ids())
1347
if text_key_references is None:
1348
text_key_references = self.find_text_key_references()
1349
pb = ui.ui_factory.nested_progress_bar()
1351
return self._do_generate_text_key_index(ancestors,
1352
text_key_references, pb)
1356
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1357
"""Helper for _generate_text_key_index to avoid deep nesting."""
1358
revision_order = tsort.topo_sort(ancestors)
1359
invalid_keys = set()
1361
for revision_id in revision_order:
1362
revision_keys[revision_id] = set()
1363
text_count = len(text_key_references)
1364
# a cache of the text keys to allow reuse; costs a dict of all the
1365
# keys, but saves a 2-tuple for every child of a given key.
1367
for text_key, valid in text_key_references.iteritems():
1369
invalid_keys.add(text_key)
1371
revision_keys[text_key[1]].add(text_key)
1372
text_key_cache[text_key] = text_key
1373
del text_key_references
1375
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1376
NULL_REVISION = _mod_revision.NULL_REVISION
1377
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1378
# too small for large or very branchy trees. However, for 55K path
1379
# trees, it would be easy to use too much memory trivially. Ideally we
1380
# could gauge this by looking at available real memory etc, but this is
1381
# always a tricky proposition.
1382
inventory_cache = lru_cache.LRUCache(10)
1383
batch_size = 10 # should be ~150MB on a 55K path tree
1384
batch_count = len(revision_order) / batch_size + 1
1386
pb.update("Calculating text parents.", processed_texts, text_count)
1387
for offset in xrange(batch_count):
1388
to_query = revision_order[offset * batch_size:(offset + 1) *
1392
for rev_tree in self.revision_trees(to_query):
1393
revision_id = rev_tree.get_revision_id()
1394
parent_ids = ancestors[revision_id]
1395
for text_key in revision_keys[revision_id]:
1396
pb.update("Calculating text parents.", processed_texts)
1397
processed_texts += 1
1398
candidate_parents = []
1399
for parent_id in parent_ids:
1400
parent_text_key = (text_key[0], parent_id)
1402
check_parent = parent_text_key not in \
1403
revision_keys[parent_id]
1405
# the parent parent_id is a ghost:
1406
check_parent = False
1407
# truncate the derived graph against this ghost.
1408
parent_text_key = None
1410
# look at the parent commit details inventories to
1411
# determine possible candidates in the per file graph.
1414
inv = inventory_cache[parent_id]
1416
inv = self.revision_tree(parent_id).inventory
1417
inventory_cache[parent_id] = inv
1418
parent_entry = inv._byid.get(text_key[0], None)
1419
if parent_entry is not None:
1421
text_key[0], parent_entry.revision)
1423
parent_text_key = None
1424
if parent_text_key is not None:
1425
candidate_parents.append(
1426
text_key_cache[parent_text_key])
1427
parent_heads = text_graph.heads(candidate_parents)
1428
new_parents = list(parent_heads)
1429
new_parents.sort(key=lambda x:candidate_parents.index(x))
1430
if new_parents == []:
1431
new_parents = [NULL_REVISION]
1432
text_index[text_key] = new_parents
1434
for text_key in invalid_keys:
1435
text_index[text_key] = [NULL_REVISION]
1438
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1439
"""Get an iterable listing the keys of all the data introduced by a set
1442
The keys will be ordered so that the corresponding items can be safely
1443
fetched and inserted in that order.
1445
:returns: An iterable producing tuples of (knit-kind, file-id,
1446
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1447
'revisions'. file-id is None unless knit-kind is 'file'.
1449
# XXX: it's a bit weird to control the inventory weave caching in this
1450
# generator. Ideally the caching would be done in fetch.py I think. Or
1451
# maybe this generator should explicitly have the contract that it
1452
# should not be iterated until the previously yielded item has been
1454
inv_w = self.get_inventory_weave()
1456
# file ids that changed
1457
file_ids = self.fileids_altered_by_revision_ids(revision_ids, inv_w)
1459
num_file_ids = len(file_ids)
1460
for file_id, altered_versions in file_ids.iteritems():
1461
if _files_pb is not None:
1462
_files_pb.update("fetch texts", count, num_file_ids)
1464
yield ("file", file_id, altered_versions)
1465
# We're done with the files_pb. Note that it finished by the caller,
1466
# just as it was created by the caller.
1470
yield ("inventory", None, revision_ids)
1473
revisions_with_signatures = set()
1474
for rev_id in revision_ids:
1476
self.get_signature_text(rev_id)
1477
except errors.NoSuchRevision:
1481
revisions_with_signatures.add(rev_id)
1482
yield ("signatures", None, revisions_with_signatures)
1485
yield ("revisions", None, revision_ids)
1488
def get_inventory_weave(self):
1489
return self.control_weaves.get_weave('inventory',
1490
self.get_transaction())
1493
def get_inventory(self, revision_id):
1494
"""Get Inventory object by revision id."""
1495
return self.iter_inventories([revision_id]).next()
1497
def iter_inventories(self, revision_ids):
1498
"""Get many inventories by revision_ids.
1500
This will buffer some or all of the texts used in constructing the
1501
inventories in memory, but will only parse a single inventory at a
1504
:return: An iterator of inventories.
1506
if ((None in revision_ids)
1507
or (_mod_revision.NULL_REVISION in revision_ids)):
1508
raise ValueError('cannot get null revision inventory')
1509
return self._iter_inventories(revision_ids)
1511
def _iter_inventories(self, revision_ids):
1512
"""single-document based inventory iteration."""
1513
texts = self.get_inventory_weave().get_texts(revision_ids)
1514
for text, revision_id in zip(texts, revision_ids):
1515
yield self.deserialise_inventory(revision_id, text)
1517
def deserialise_inventory(self, revision_id, xml):
1518
"""Transform the xml into an inventory object.
1520
:param revision_id: The expected revision id of the inventory.
1521
:param xml: A serialised inventory.
1523
result = self._serializer.read_inventory_from_string(xml, revision_id)
1524
if result.revision_id != revision_id:
1525
raise AssertionError('revision id mismatch %s != %s' % (
1526
result.revision_id, revision_id))
1529
def serialise_inventory(self, inv):
1530
return self._serializer.write_inventory_to_string(inv)
1532
def _serialise_inventory_to_lines(self, inv):
1533
return self._serializer.write_inventory_to_lines(inv)
1535
def get_serializer_format(self):
1536
return self._serializer.format_num
1539
def get_inventory_xml(self, revision_id):
1540
"""Get inventory XML as a file object."""
1542
iw = self.get_inventory_weave()
1543
return iw.get_text(revision_id)
1545
raise errors.HistoryMissing(self, 'inventory', revision_id)
1548
def get_inventory_sha1(self, revision_id):
1549
"""Return the sha1 hash of the inventory entry
1551
return self.get_revision(revision_id).inventory_sha1
1553
def iter_reverse_revision_history(self, revision_id):
1554
"""Iterate backwards through revision ids in the lefthand history
1556
:param revision_id: The revision id to start with. All its lefthand
1557
ancestors will be traversed.
1559
graph = self.get_graph()
1560
next_id = revision_id
1562
if next_id in (None, _mod_revision.NULL_REVISION):
1565
# Note: The following line may raise KeyError in the event of
1566
# truncated history. We decided not to have a try:except:raise
1567
# RevisionNotPresent here until we see a use for it, because of the
1568
# cost in an inner loop that is by its very nature O(history).
1569
# Robert Collins 20080326
1570
parents = graph.get_parent_map([next_id])[next_id]
1571
if len(parents) == 0:
1574
next_id = parents[0]
1577
def get_revision_inventory(self, revision_id):
1578
"""Return inventory of a past revision."""
1579
# TODO: Unify this with get_inventory()
1580
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1581
# must be the same as its revision, so this is trivial.
1582
if revision_id is None:
1583
# This does not make sense: if there is no revision,
1584
# then it is the current tree inventory surely ?!
1585
# and thus get_root_id() is something that looks at the last
1586
# commit on the branch, and the get_root_id is an inventory check.
1587
raise NotImplementedError
1588
# return Inventory(self.get_root_id())
1590
return self.get_inventory(revision_id)
1593
def is_shared(self):
1594
"""Return True if this repository is flagged as a shared repository."""
1595
raise NotImplementedError(self.is_shared)
1598
def reconcile(self, other=None, thorough=False):
1599
"""Reconcile this repository."""
1600
from bzrlib.reconcile import RepoReconciler
1601
reconciler = RepoReconciler(self, thorough=thorough)
1602
reconciler.reconcile()
1605
def _refresh_data(self):
1606
"""Helper called from lock_* to ensure coherency with disk.
1608
The default implementation does nothing; it is however possible
1609
for repositories to maintain loaded indices across multiple locks
1610
by checking inside their implementation of this method to see
1611
whether their indices are still valid. This depends of course on
1612
the disk format being validatable in this manner.
1616
def revision_tree(self, revision_id):
1617
"""Return Tree for a revision on this branch.
1619
`revision_id` may be None for the empty tree revision.
1621
# TODO: refactor this to use an existing revision object
1622
# so we don't need to read it in twice.
1623
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1624
return RevisionTree(self, Inventory(root_id=None),
1625
_mod_revision.NULL_REVISION)
1627
inv = self.get_revision_inventory(revision_id)
1628
return RevisionTree(self, inv, revision_id)
1630
def revision_trees(self, revision_ids):
1631
"""Return Tree for a revision on this branch.
1633
`revision_id` may not be None or 'null:'"""
1634
inventories = self.iter_inventories(revision_ids)
1635
for inv in inventories:
1636
yield RevisionTree(self, inv, inv.revision_id)
1639
def get_ancestry(self, revision_id, topo_sorted=True):
1640
"""Return a list of revision-ids integrated by a revision.
1642
The first element of the list is always None, indicating the origin
1643
revision. This might change when we have history horizons, or
1644
perhaps we should have a new API.
1646
This is topologically sorted.
1648
if _mod_revision.is_null(revision_id):
1650
if not self.has_revision(revision_id):
1651
raise errors.NoSuchRevision(self, revision_id)
1652
w = self.get_inventory_weave()
1653
candidates = w.get_ancestry(revision_id, topo_sorted)
1654
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1657
"""Compress the data within the repository.
1659
This operation only makes sense for some repository types. For other
1660
types it should be a no-op that just returns.
1662
This stub method does not require a lock, but subclasses should use
1663
@needs_write_lock as this is a long running call its reasonable to
1664
implicitly lock for the user.
1668
def print_file(self, file, revision_id):
1669
"""Print `file` to stdout.
1671
FIXME RBC 20060125 as John Meinel points out this is a bad api
1672
- it writes to stdout, it assumes that that is valid etc. Fix
1673
by creating a new more flexible convenience function.
1675
tree = self.revision_tree(revision_id)
1676
# use inventory as it was in that revision
1677
file_id = tree.inventory.path2id(file)
1679
# TODO: jam 20060427 Write a test for this code path
1680
# it had a bug in it, and was raising the wrong
1682
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1683
tree.print_file(file_id)
1685
def get_transaction(self):
1686
return self.control_files.get_transaction()
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
:ivar _transport: Transport for access to repository control files,
1944
typically pointing to .bzr/repository.
1947
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1948
super(MetaDirRepository, self).__init__(_format,
1954
self._transport = control_files._transport
1957
def is_shared(self):
1958
"""Return True if this repository is flagged as a shared repository."""
1959
return self._transport.has('shared-storage')
1962
def set_make_working_trees(self, new_value):
1963
"""Set the policy flag for making working trees when creating branches.
1965
This only applies to branches that use this repository.
1967
The default is 'True'.
1968
:param new_value: True to restore the default, False to disable making
1973
self._transport.delete('no-working-trees')
1974
except errors.NoSuchFile:
1977
self._transport.put_bytes('no-working-trees', '',
1978
mode=self.bzrdir._get_file_mode())
1980
def make_working_trees(self):
1981
"""Returns the policy for making working trees on new branches."""
1982
return not self._transport.has('no-working-trees')
1985
class MetaDirVersionedFileRepository(MetaDirRepository):
1986
"""Repositories in a meta-dir, that work via versioned file objects."""
1988
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1989
super(MetaDirVersionedFileRepository, self).__init__(_format, a_bzrdir,
1990
control_files, _revision_store, control_store, text_store)
1991
_revision_store.get_scope = self.get_transaction
1992
control_store.get_scope = self.get_transaction
1993
text_store.get_scope = self.get_transaction
1996
class RepositoryFormatRegistry(registry.Registry):
1997
"""Registry of RepositoryFormats."""
1999
def get(self, format_string):
2000
r = registry.Registry.get(self, format_string)
2006
format_registry = RepositoryFormatRegistry()
2007
"""Registry of formats, indexed by their identifying format string.
2009
This can contain either format instances themselves, or classes/factories that
2010
can be called to obtain one.
2014
#####################################################################
2015
# Repository Formats
2017
class RepositoryFormat(object):
2018
"""A repository format.
2020
Formats provide three things:
2021
* An initialization routine to construct repository data on disk.
2022
* a format string which is used when the BzrDir supports versioned
2024
* an open routine which returns a Repository instance.
2026
There is one and only one Format subclass for each on-disk format. But
2027
there can be one Repository subclass that is used for several different
2028
formats. The _format attribute on a Repository instance can be used to
2029
determine the disk format.
2031
Formats are placed in an dict by their format string for reference
2032
during opening. These should be subclasses of RepositoryFormat
2035
Once a format is deprecated, just deprecate the initialize and open
2036
methods on the format class. Do not deprecate the object, as the
2037
object will be created every system load.
2039
Common instance attributes:
2040
_matchingbzrdir - the bzrdir format that the repository format was
2041
originally written to work with. This can be used if manually
2042
constructing a bzrdir and repository, or more commonly for test suite
2046
# Set to True or False in derived classes. True indicates that the format
2047
# supports ghosts gracefully.
2048
supports_ghosts = None
2049
# Can this repository be given external locations to lookup additional
2050
# data. Set to True or False in derived classes.
2051
supports_external_lookups = None
2054
return "<%s>" % self.__class__.__name__
2056
def __eq__(self, other):
2057
# format objects are generally stateless
2058
return isinstance(other, self.__class__)
2060
def __ne__(self, other):
2061
return not self == other
2064
def find_format(klass, a_bzrdir):
2065
"""Return the format for the repository object in a_bzrdir.
2067
This is used by bzr native formats that have a "format" file in
2068
the repository. Other methods may be used by different types of
2072
transport = a_bzrdir.get_repository_transport(None)
2073
format_string = transport.get("format").read()
2074
return format_registry.get(format_string)
2075
except errors.NoSuchFile:
2076
raise errors.NoRepositoryPresent(a_bzrdir)
2078
raise errors.UnknownFormatError(format=format_string,
2082
def register_format(klass, format):
2083
format_registry.register(format.get_format_string(), format)
2086
def unregister_format(klass, format):
2087
format_registry.remove(format.get_format_string())
2090
def get_default_format(klass):
2091
"""Return the current default format."""
2092
from bzrlib import bzrdir
2093
return bzrdir.format_registry.make_bzrdir('default').repository_format
2095
def _get_control_store(self, repo_transport, control_files):
2096
"""Return the control store for this repository."""
2097
raise NotImplementedError(self._get_control_store)
2099
def get_format_string(self):
2100
"""Return the ASCII format string that identifies this format.
2102
Note that in pre format ?? repositories the format string is
2103
not permitted nor written to disk.
2105
raise NotImplementedError(self.get_format_string)
2107
def get_format_description(self):
2108
"""Return the short description for this format."""
2109
raise NotImplementedError(self.get_format_description)
2111
def _get_revision_store(self, repo_transport, control_files):
2112
"""Return the revision store object for this a_bzrdir."""
2113
raise NotImplementedError(self._get_revision_store)
2115
def _get_text_rev_store(self,
2122
"""Common logic for getting a revision store for a repository.
2124
see self._get_revision_store for the subclass-overridable method to
2125
get the store for a repository.
2127
from bzrlib.store.revision.text import TextRevisionStore
2128
dir_mode = control_files._dir_mode
2129
file_mode = control_files._file_mode
2130
text_store = TextStore(transport.clone(name),
2132
compressed=compressed,
2134
file_mode=file_mode)
2135
_revision_store = TextRevisionStore(text_store, serializer)
2136
return _revision_store
2138
# TODO: this shouldn't be in the base class, it's specific to things that
2139
# use weaves or knits -- mbp 20070207
2140
def _get_versioned_file_store(self,
2145
versionedfile_class=None,
2146
versionedfile_kwargs={},
2148
if versionedfile_class is None:
2149
versionedfile_class = self._versionedfile_class
2150
weave_transport = control_files._transport.clone(name)
2151
dir_mode = control_files._dir_mode
2152
file_mode = control_files._file_mode
2153
return VersionedFileStore(weave_transport, prefixed=prefixed,
2155
file_mode=file_mode,
2156
versionedfile_class=versionedfile_class,
2157
versionedfile_kwargs=versionedfile_kwargs,
2160
def initialize(self, a_bzrdir, shared=False):
2161
"""Initialize a repository of this format in a_bzrdir.
2163
:param a_bzrdir: The bzrdir to put the new repository in it.
2164
:param shared: The repository should be initialized as a sharable one.
2165
:returns: The new repository object.
2167
This may raise UninitializableFormat if shared repository are not
2168
compatible the a_bzrdir.
2170
raise NotImplementedError(self.initialize)
2172
def is_supported(self):
2173
"""Is this format supported?
2175
Supported formats must be initializable and openable.
2176
Unsupported formats may not support initialization or committing or
2177
some other features depending on the reason for not being supported.
2181
def check_conversion_target(self, target_format):
2182
raise NotImplementedError(self.check_conversion_target)
2184
def open(self, a_bzrdir, _found=False):
2185
"""Return an instance of this format for the bzrdir a_bzrdir.
2187
_found is a private parameter, do not use it.
2189
raise NotImplementedError(self.open)
2192
class MetaDirRepositoryFormat(RepositoryFormat):
2193
"""Common base class for the new repositories using the metadir layout."""
2195
rich_root_data = False
2196
supports_tree_reference = False
2197
supports_external_lookups = False
2198
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
2201
super(MetaDirRepositoryFormat, self).__init__()
2203
def _create_control_files(self, a_bzrdir):
2204
"""Create the required files and the initial control_files object."""
2205
# FIXME: RBC 20060125 don't peek under the covers
2206
# NB: no need to escape relative paths that are url safe.
2207
repository_transport = a_bzrdir.get_repository_transport(self)
2208
control_files = lockable_files.LockableFiles(repository_transport,
2209
'lock', lockdir.LockDir)
2210
control_files.create_lock()
2211
return control_files
2213
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
2214
"""Upload the initial blank content."""
2215
control_files = self._create_control_files(a_bzrdir)
2216
control_files.lock_write()
2217
transport = control_files._transport
2219
utf8_files += [('shared-storage', '')]
2221
transport.mkdir_multi(dirs, mode=a_bzrdir._get_dir_mode())
2222
for (filename, content_stream) in files:
2223
transport.put_file(filename, content_stream,
2224
mode=a_bzrdir._get_file_mode())
2225
for (filename, content_bytes) in utf8_files:
2226
transport.put_bytes_non_atomic(filename, content_bytes,
2227
mode=a_bzrdir._get_file_mode())
2229
control_files.unlock()
2232
# formats which have no format string are not discoverable
2233
# and not independently creatable, so are not registered. They're
2234
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
2235
# needed, it's constructed directly by the BzrDir. Non-native formats where
2236
# the repository is not separately opened are similar.
2238
format_registry.register_lazy(
2239
'Bazaar-NG Repository format 7',
2240
'bzrlib.repofmt.weaverepo',
2244
format_registry.register_lazy(
2245
'Bazaar-NG Knit Repository Format 1',
2246
'bzrlib.repofmt.knitrepo',
2247
'RepositoryFormatKnit1',
2250
format_registry.register_lazy(
2251
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
2252
'bzrlib.repofmt.knitrepo',
2253
'RepositoryFormatKnit3',
2256
format_registry.register_lazy(
2257
'Bazaar Knit Repository Format 4 (bzr 1.0)\n',
2258
'bzrlib.repofmt.knitrepo',
2259
'RepositoryFormatKnit4',
2262
# Pack-based formats. There is one format for pre-subtrees, and one for
2263
# post-subtrees to allow ease of testing.
2264
# NOTE: These are experimental in 0.92. Stable in 1.0 and above
2265
format_registry.register_lazy(
2266
'Bazaar pack repository format 1 (needs bzr 0.92)\n',
2267
'bzrlib.repofmt.pack_repo',
2268
'RepositoryFormatKnitPack1',
2270
format_registry.register_lazy(
2271
'Bazaar pack repository format 1 with subtree support (needs bzr 0.92)\n',
2272
'bzrlib.repofmt.pack_repo',
2273
'RepositoryFormatKnitPack3',
2275
format_registry.register_lazy(
2276
'Bazaar pack repository format 1 with rich root (needs bzr 1.0)\n',
2277
'bzrlib.repofmt.pack_repo',
2278
'RepositoryFormatKnitPack4',
2280
# Development formats.
2282
# development 0 - stub to introduce development versioning scheme.
2283
format_registry.register_lazy(
2284
"Bazaar development format 0 (needs bzr.dev from before 1.3)\n",
2285
'bzrlib.repofmt.pack_repo',
2286
'RepositoryFormatPackDevelopment0',
2288
format_registry.register_lazy(
2289
("Bazaar development format 0 with subtree support "
2290
"(needs bzr.dev from before 1.3)\n"),
2291
'bzrlib.repofmt.pack_repo',
2292
'RepositoryFormatPackDevelopment0Subtree',
2294
# 1.3->1.4 go below here
2297
class InterRepository(InterObject):
2298
"""This class represents operations taking place between two repositories.
2300
Its instances have methods like copy_content and fetch, and contain
2301
references to the source and target repositories these operations can be
2304
Often we will provide convenience methods on 'repository' which carry out
2305
operations with another repository - they will always forward to
2306
InterRepository.get(other).method_name(parameters).
2310
"""The available optimised InterRepository types."""
2312
def copy_content(self, revision_id=None):
2313
raise NotImplementedError(self.copy_content)
2315
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2316
"""Fetch the content required to construct revision_id.
2318
The content is copied from self.source to self.target.
2320
:param revision_id: if None all content is copied, if NULL_REVISION no
2322
:param pb: optional progress bar to use for progress reports. If not
2323
provided a default one will be created.
2325
Returns the copied revision count and the failed revisions in a tuple:
2328
raise NotImplementedError(self.fetch)
2330
def _walk_to_common_revisions(self, revision_ids):
2331
"""Walk out from revision_ids in source to revisions target has.
2333
:param revision_ids: The start point for the search.
2334
:return: A set of revision ids.
2336
target_graph = self.target.get_graph()
2337
revision_ids = frozenset(revision_ids)
2338
if set(target_graph.get_parent_map(revision_ids)) == revision_ids:
2339
return graph.SearchResult(revision_ids, set(), 0, set())
2340
missing_revs = set()
2341
source_graph = self.source.get_graph()
2342
# ensure we don't pay silly lookup costs.
2343
searcher = source_graph._make_breadth_first_searcher(revision_ids)
2344
null_set = frozenset([_mod_revision.NULL_REVISION])
2347
next_revs, ghosts = searcher.next_with_ghosts()
2348
except StopIteration:
2350
if revision_ids.intersection(ghosts):
2351
absent_ids = set(revision_ids.intersection(ghosts))
2352
# If all absent_ids are present in target, no error is needed.
2353
absent_ids.difference_update(
2354
set(target_graph.get_parent_map(absent_ids)))
2356
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2357
# we don't care about other ghosts as we can't fetch them and
2358
# haven't been asked to.
2359
next_revs = set(next_revs)
2360
# we always have NULL_REVISION present.
2361
have_revs = set(target_graph.get_parent_map(next_revs)).union(null_set)
2362
missing_revs.update(next_revs - have_revs)
2363
searcher.stop_searching_any(have_revs)
2364
return searcher.get_result()
2366
@deprecated_method(symbol_versioning.one_two)
2368
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2369
"""Return the revision ids that source has that target does not.
2371
These are returned in topological order.
2373
:param revision_id: only return revision ids included by this
2375
:param find_ghosts: If True find missing revisions in deep history
2376
rather than just finding the surface difference.
2378
return list(self.search_missing_revision_ids(
2379
revision_id, find_ghosts).get_keys())
2382
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2383
"""Return the revision ids that source has that target does not.
2385
:param revision_id: only return revision ids included by this
2387
:param find_ghosts: If True find missing revisions in deep history
2388
rather than just finding the surface difference.
2389
:return: A bzrlib.graph.SearchResult.
2391
# stop searching at found target revisions.
2392
if not find_ghosts and revision_id is not None:
2393
return self._walk_to_common_revisions([revision_id])
2394
# generic, possibly worst case, slow code path.
2395
target_ids = set(self.target.all_revision_ids())
2396
if revision_id is not None:
2397
source_ids = self.source.get_ancestry(revision_id)
2398
if source_ids[0] is not None:
2399
raise AssertionError()
2402
source_ids = self.source.all_revision_ids()
2403
result_set = set(source_ids).difference(target_ids)
2404
return self.source.revision_ids_to_search_result(result_set)
2407
def _same_model(source, target):
2408
"""True if source and target have the same data representation."""
2409
if source.supports_rich_root() != target.supports_rich_root():
2411
if source._serializer != target._serializer:
2416
class InterSameDataRepository(InterRepository):
2417
"""Code for converting between repositories that represent the same data.
2419
Data format and model must match for this to work.
2423
def _get_repo_format_to_test(self):
2424
"""Repository format for testing with.
2426
InterSameData can pull from subtree to subtree and from non-subtree to
2427
non-subtree, so we test this with the richest repository format.
2429
from bzrlib.repofmt import knitrepo
2430
return knitrepo.RepositoryFormatKnit3()
2433
def is_compatible(source, target):
2434
return InterRepository._same_model(source, target)
2437
def copy_content(self, revision_id=None):
2438
"""Make a complete copy of the content in self into destination.
2440
This copies both the repository's revision data, and configuration information
2441
such as the make_working_trees setting.
2443
This is a destructive operation! Do not use it on existing
2446
:param revision_id: Only copy the content needed to construct
2447
revision_id and its parents.
2450
self.target.set_make_working_trees(self.source.make_working_trees())
2451
except NotImplementedError:
2453
# but don't bother fetching if we have the needed data now.
2454
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2455
self.target.has_revision(revision_id)):
2457
self.target.fetch(self.source, revision_id=revision_id)
2460
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2461
"""See InterRepository.fetch()."""
2462
from bzrlib.fetch import GenericRepoFetcher
2463
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2464
self.source, self.source._format, self.target,
2465
self.target._format)
2466
f = GenericRepoFetcher(to_repository=self.target,
2467
from_repository=self.source,
2468
last_revision=revision_id,
2469
pb=pb, find_ghosts=find_ghosts)
2470
return f.count_copied, f.failed_revisions
2473
class InterWeaveRepo(InterSameDataRepository):
2474
"""Optimised code paths between Weave based repositories.
2476
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2477
implemented lazy inter-object optimisation.
2481
def _get_repo_format_to_test(self):
2482
from bzrlib.repofmt import weaverepo
2483
return weaverepo.RepositoryFormat7()
2486
def is_compatible(source, target):
2487
"""Be compatible with known Weave formats.
2489
We don't test for the stores being of specific types because that
2490
could lead to confusing results, and there is no need to be
2493
from bzrlib.repofmt.weaverepo import (
2499
return (isinstance(source._format, (RepositoryFormat5,
2501
RepositoryFormat7)) and
2502
isinstance(target._format, (RepositoryFormat5,
2504
RepositoryFormat7)))
2505
except AttributeError:
2509
def copy_content(self, revision_id=None):
2510
"""See InterRepository.copy_content()."""
2511
# weave specific optimised path:
2513
self.target.set_make_working_trees(self.source.make_working_trees())
2514
except (errors.RepositoryUpgradeRequired, NotImplemented):
2516
# FIXME do not peek!
2517
if self.source._transport.listable():
2518
pb = ui.ui_factory.nested_progress_bar()
2520
self.target.weave_store.copy_all_ids(
2521
self.source.weave_store,
2523
from_transaction=self.source.get_transaction(),
2524
to_transaction=self.target.get_transaction())
2525
pb.update('copying inventory', 0, 1)
2526
self.target.control_weaves.copy_multi(
2527
self.source.control_weaves, ['inventory'],
2528
from_transaction=self.source.get_transaction(),
2529
to_transaction=self.target.get_transaction())
2530
self.target._revision_store.text_store.copy_all_ids(
2531
self.source._revision_store.text_store,
2536
self.target.fetch(self.source, revision_id=revision_id)
2539
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2540
"""See InterRepository.fetch()."""
2541
from bzrlib.fetch import GenericRepoFetcher
2542
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2543
self.source, self.source._format, self.target, self.target._format)
2544
f = GenericRepoFetcher(to_repository=self.target,
2545
from_repository=self.source,
2546
last_revision=revision_id,
2547
pb=pb, find_ghosts=find_ghosts)
2548
return f.count_copied, f.failed_revisions
2551
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2552
"""See InterRepository.missing_revision_ids()."""
2553
# we want all revisions to satisfy revision_id in source.
2554
# but we don't want to stat every file here and there.
2555
# we want then, all revisions other needs to satisfy revision_id
2556
# checked, but not those that we have locally.
2557
# so the first thing is to get a subset of the revisions to
2558
# satisfy revision_id in source, and then eliminate those that
2559
# we do already have.
2560
# this is slow on high latency connection to self, but as as this
2561
# disk format scales terribly for push anyway due to rewriting
2562
# inventory.weave, this is considered acceptable.
2564
if revision_id is not None:
2565
source_ids = self.source.get_ancestry(revision_id)
2566
if source_ids[0] is not None:
2567
raise AssertionError()
2570
source_ids = self.source._all_possible_ids()
2571
source_ids_set = set(source_ids)
2572
# source_ids is the worst possible case we may need to pull.
2573
# now we want to filter source_ids against what we actually
2574
# have in target, but don't try to check for existence where we know
2575
# we do not have a revision as that would be pointless.
2576
target_ids = set(self.target._all_possible_ids())
2577
possibly_present_revisions = target_ids.intersection(source_ids_set)
2578
actually_present_revisions = set(
2579
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2580
required_revisions = source_ids_set.difference(actually_present_revisions)
2581
if revision_id is not None:
2582
# we used get_ancestry to determine source_ids then we are assured all
2583
# revisions referenced are present as they are installed in topological order.
2584
# and the tip revision was validated by get_ancestry.
2585
result_set = required_revisions
2587
# if we just grabbed the possibly available ids, then
2588
# we only have an estimate of whats available and need to validate
2589
# that against the revision records.
2591
self.source._eliminate_revisions_not_present(required_revisions))
2592
return self.source.revision_ids_to_search_result(result_set)
2595
class InterKnitRepo(InterSameDataRepository):
2596
"""Optimised code paths between Knit based repositories."""
2599
def _get_repo_format_to_test(self):
2600
from bzrlib.repofmt import knitrepo
2601
return knitrepo.RepositoryFormatKnit1()
2604
def is_compatible(source, target):
2605
"""Be compatible with known Knit formats.
2607
We don't test for the stores being of specific types because that
2608
could lead to confusing results, and there is no need to be
2611
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2613
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2614
isinstance(target._format, RepositoryFormatKnit))
2615
except AttributeError:
2617
return are_knits and InterRepository._same_model(source, target)
2620
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2621
"""See InterRepository.fetch()."""
2622
from bzrlib.fetch import KnitRepoFetcher
2623
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2624
self.source, self.source._format, self.target, self.target._format)
2625
f = KnitRepoFetcher(to_repository=self.target,
2626
from_repository=self.source,
2627
last_revision=revision_id,
2628
pb=pb, find_ghosts=find_ghosts)
2629
return f.count_copied, f.failed_revisions
2632
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2633
"""See InterRepository.missing_revision_ids()."""
2634
if revision_id is not None:
2635
source_ids = self.source.get_ancestry(revision_id)
2636
if source_ids[0] is not None:
2637
raise AssertionError()
2640
source_ids = self.source.all_revision_ids()
2641
source_ids_set = set(source_ids)
2642
# source_ids is the worst possible case we may need to pull.
2643
# now we want to filter source_ids against what we actually
2644
# have in target, but don't try to check for existence where we know
2645
# we do not have a revision as that would be pointless.
2646
target_ids = set(self.target.all_revision_ids())
2647
possibly_present_revisions = target_ids.intersection(source_ids_set)
2648
actually_present_revisions = set(
2649
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2650
required_revisions = source_ids_set.difference(actually_present_revisions)
2651
if revision_id is not None:
2652
# we used get_ancestry to determine source_ids then we are assured all
2653
# revisions referenced are present as they are installed in topological order.
2654
# and the tip revision was validated by get_ancestry.
2655
result_set = required_revisions
2657
# if we just grabbed the possibly available ids, then
2658
# we only have an estimate of whats available and need to validate
2659
# that against the revision records.
2661
self.source._eliminate_revisions_not_present(required_revisions))
2662
return self.source.revision_ids_to_search_result(result_set)
2665
class InterPackRepo(InterSameDataRepository):
2666
"""Optimised code paths between Pack based repositories."""
2669
def _get_repo_format_to_test(self):
2670
from bzrlib.repofmt import pack_repo
2671
return pack_repo.RepositoryFormatKnitPack1()
2674
def is_compatible(source, target):
2675
"""Be compatible with known Pack formats.
2677
We don't test for the stores being of specific types because that
2678
could lead to confusing results, and there is no need to be
2681
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2683
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2684
isinstance(target._format, RepositoryFormatPack))
2685
except AttributeError:
2687
return are_packs and InterRepository._same_model(source, target)
2690
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2691
"""See InterRepository.fetch()."""
2692
from bzrlib.repofmt.pack_repo import Packer
2693
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2694
self.source, self.source._format, self.target, self.target._format)
2695
self.count_copied = 0
2696
if revision_id is None:
2698
# everything to do - use pack logic
2699
# to fetch from all packs to one without
2700
# inventory parsing etc, IFF nothing to be copied is in the target.
2702
revision_ids = self.source.all_revision_ids()
2703
revision_keys = [(revid,) for revid in revision_ids]
2704
index = self.target._pack_collection.revision_index.combined_index
2705
present_revision_ids = set(item[1][0] for item in
2706
index.iter_entries(revision_keys))
2707
revision_ids = set(revision_ids) - present_revision_ids
2708
# implementing the TODO will involve:
2709
# - detecting when all of a pack is selected
2710
# - avoiding as much as possible pre-selection, so the
2711
# more-core routines such as create_pack_from_packs can filter in
2712
# a just-in-time fashion. (though having a HEADS list on a
2713
# repository might make this a lot easier, because we could
2714
# sensibly detect 'new revisions' without doing a full index scan.
2715
elif _mod_revision.is_null(revision_id):
2720
revision_ids = self.search_missing_revision_ids(revision_id,
2721
find_ghosts=find_ghosts).get_keys()
2722
except errors.NoSuchRevision:
2723
raise errors.InstallFailed([revision_id])
2724
if len(revision_ids) == 0:
2726
packs = self.source._pack_collection.all_packs()
2727
pack = Packer(self.target._pack_collection, packs, '.fetch',
2728
revision_ids).pack()
2729
if pack is not None:
2730
self.target._pack_collection._save_pack_names()
2731
# Trigger an autopack. This may duplicate effort as we've just done
2732
# a pack creation, but for now it is simpler to think about as
2733
# 'upload data, then repack if needed'.
2734
self.target._pack_collection.autopack()
2735
return (pack.get_revision_count(), [])
2740
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2741
"""See InterRepository.missing_revision_ids().
2743
:param find_ghosts: Find ghosts throughout the ancestry of
2746
if not find_ghosts and revision_id is not None:
2747
return self._walk_to_common_revisions([revision_id])
2748
elif revision_id is not None:
2749
source_ids = self.source.get_ancestry(revision_id)
2750
if source_ids[0] is not None:
2751
raise AssertionError()
2754
source_ids = self.source.all_revision_ids()
2755
# source_ids is the worst possible case we may need to pull.
2756
# now we want to filter source_ids against what we actually
2757
# have in target, but don't try to check for existence where we know
2758
# we do not have a revision as that would be pointless.
2759
target_ids = set(self.target.all_revision_ids())
2760
result_set = set(source_ids).difference(target_ids)
2761
return self.source.revision_ids_to_search_result(result_set)
2764
class InterModel1and2(InterRepository):
2767
def _get_repo_format_to_test(self):
2771
def is_compatible(source, target):
2772
if not source.supports_rich_root() and target.supports_rich_root():
2778
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2779
"""See InterRepository.fetch()."""
2780
from bzrlib.fetch import Model1toKnit2Fetcher
2781
f = Model1toKnit2Fetcher(to_repository=self.target,
2782
from_repository=self.source,
2783
last_revision=revision_id,
2784
pb=pb, find_ghosts=find_ghosts)
2785
return f.count_copied, f.failed_revisions
2788
def copy_content(self, revision_id=None):
2789
"""Make a complete copy of the content in self into destination.
2791
This is a destructive operation! Do not use it on existing
2794
:param revision_id: Only copy the content needed to construct
2795
revision_id and its parents.
2798
self.target.set_make_working_trees(self.source.make_working_trees())
2799
except NotImplementedError:
2801
# but don't bother fetching if we have the needed data now.
2802
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2803
self.target.has_revision(revision_id)):
2805
self.target.fetch(self.source, revision_id=revision_id)
2808
class InterKnit1and2(InterKnitRepo):
2811
def _get_repo_format_to_test(self):
2815
def is_compatible(source, target):
2816
"""Be compatible with Knit1 source and Knit3 target"""
2817
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2819
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2820
RepositoryFormatKnit3)
2821
from bzrlib.repofmt.pack_repo import (
2822
RepositoryFormatKnitPack1,
2823
RepositoryFormatKnitPack3,
2824
RepositoryFormatPackDevelopment0,
2825
RepositoryFormatPackDevelopment0Subtree,
2828
RepositoryFormatKnit1,
2829
RepositoryFormatKnitPack1,
2830
RepositoryFormatPackDevelopment0,
2833
RepositoryFormatKnit3,
2834
RepositoryFormatKnitPack3,
2835
RepositoryFormatPackDevelopment0Subtree,
2837
return (isinstance(source._format, nosubtrees) and
2838
isinstance(target._format, subtrees))
2839
except AttributeError:
2843
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2844
"""See InterRepository.fetch()."""
2845
from bzrlib.fetch import Knit1to2Fetcher
2846
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2847
self.source, self.source._format, self.target,
2848
self.target._format)
2849
f = Knit1to2Fetcher(to_repository=self.target,
2850
from_repository=self.source,
2851
last_revision=revision_id,
2852
pb=pb, find_ghosts=find_ghosts)
2853
return f.count_copied, f.failed_revisions
2856
class InterDifferingSerializer(InterKnitRepo):
2859
def _get_repo_format_to_test(self):
2863
def is_compatible(source, target):
2864
"""Be compatible with Knit2 source and Knit3 target"""
2865
if source.supports_rich_root() != target.supports_rich_root():
2867
# Ideally, we'd support fetching if the source had no tree references
2868
# even if it supported them...
2869
if (getattr(source, '_format.supports_tree_reference', False) and
2870
not getattr(target, '_format.supports_tree_reference', False)):
2875
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2876
"""See InterRepository.fetch()."""
2877
revision_ids = self.target.search_missing_revision_ids(self.source,
2878
revision_id, find_ghosts=find_ghosts).get_keys()
2879
revision_ids = tsort.topo_sort(
2880
self.source.get_graph().get_parent_map(revision_ids))
2881
def revisions_iterator():
2882
for current_revision_id in revision_ids:
2883
revision = self.source.get_revision(current_revision_id)
2884
tree = self.source.revision_tree(current_revision_id)
2886
signature = self.source.get_signature_text(
2887
current_revision_id)
2888
except errors.NoSuchRevision:
2890
yield revision, tree, signature
2892
my_pb = ui.ui_factory.nested_progress_bar()
2897
install_revisions(self.target, revisions_iterator(),
2898
len(revision_ids), pb)
2900
if my_pb is not None:
2902
return len(revision_ids), 0
2905
class InterRemoteToOther(InterRepository):
2907
def __init__(self, source, target):
2908
InterRepository.__init__(self, source, target)
2909
self._real_inter = None
2912
def is_compatible(source, target):
2913
if not isinstance(source, remote.RemoteRepository):
2915
# Is source's model compatible with target's model?
2916
source._ensure_real()
2917
real_source = source._real_repository
2918
if isinstance(real_source, remote.RemoteRepository):
2919
raise NotImplementedError(
2920
"We don't support remote repos backed by remote repos yet.")
2921
return InterRepository._same_model(real_source, target)
2924
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2925
"""See InterRepository.fetch()."""
2926
from bzrlib.fetch import RemoteToOtherFetcher
2927
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2928
self.source, self.target, self.target._format)
2929
# TODO: jam 20070210 This should be an assert, not a translate
2930
revision_id = osutils.safe_revision_id(revision_id)
2931
f = RemoteToOtherFetcher(to_repository=self.target,
2932
from_repository=self.source,
2933
last_revision=revision_id,
2934
pb=pb, find_ghosts=find_ghosts)
2935
return f.count_copied, f.failed_revisions
2938
def _get_repo_format_to_test(self):
2942
class InterOtherToRemote(InterRepository):
2944
def __init__(self, source, target):
2945
InterRepository.__init__(self, source, target)
2946
self._real_inter = None
2949
def is_compatible(source, target):
2950
if isinstance(target, remote.RemoteRepository):
2954
def _ensure_real_inter(self):
2955
if self._real_inter is None:
2956
self.target._ensure_real()
2957
real_target = self.target._real_repository
2958
self._real_inter = InterRepository.get(self.source, real_target)
2960
def copy_content(self, revision_id=None):
2961
self._ensure_real_inter()
2962
self._real_inter.copy_content(revision_id=revision_id)
2964
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2965
self._ensure_real_inter()
2966
self._real_inter.fetch(revision_id=revision_id, pb=pb,
2967
find_ghosts=find_ghosts)
2970
def _get_repo_format_to_test(self):
2974
InterRepository.register_optimiser(InterDifferingSerializer)
2975
InterRepository.register_optimiser(InterSameDataRepository)
2976
InterRepository.register_optimiser(InterWeaveRepo)
2977
InterRepository.register_optimiser(InterKnitRepo)
2978
InterRepository.register_optimiser(InterModel1and2)
2979
InterRepository.register_optimiser(InterKnit1and2)
2980
InterRepository.register_optimiser(InterPackRepo)
2981
InterRepository.register_optimiser(InterRemoteToOther)
2982
InterRepository.register_optimiser(InterOtherToRemote)
2985
class CopyConverter(object):
2986
"""A repository conversion tool which just performs a copy of the content.
2988
This is slow but quite reliable.
2991
def __init__(self, target_format):
2992
"""Create a CopyConverter.
2994
:param target_format: The format the resulting repository should be.
2996
self.target_format = target_format
2998
def convert(self, repo, pb):
2999
"""Perform the conversion of to_convert, giving feedback via pb.
3001
:param to_convert: The disk object to convert.
3002
:param pb: a progress bar to use for progress information.
3007
# this is only useful with metadir layouts - separated repo content.
3008
# trigger an assertion if not such
3009
repo._format.get_format_string()
3010
self.repo_dir = repo.bzrdir
3011
self.step('Moving repository to repository.backup')
3012
self.repo_dir.transport.move('repository', 'repository.backup')
3013
backup_transport = self.repo_dir.transport.clone('repository.backup')
3014
repo._format.check_conversion_target(self.target_format)
3015
self.source_repo = repo._format.open(self.repo_dir,
3017
_override_transport=backup_transport)
3018
self.step('Creating new repository')
3019
converted = self.target_format.initialize(self.repo_dir,
3020
self.source_repo.is_shared())
3021
converted.lock_write()
3023
self.step('Copying content into repository.')
3024
self.source_repo.copy_content_into(converted)
3027
self.step('Deleting old repository content.')
3028
self.repo_dir.transport.delete_tree('repository.backup')
3029
self.pb.note('repository converted')
3031
def step(self, message):
3032
"""Update the pb by a step."""
3034
self.pb.update(message, self.count, self.total)
3046
def _unescaper(match, _map=_unescape_map):
3047
code = match.group(1)
3051
if not code.startswith('#'):
3053
return unichr(int(code[1:])).encode('utf8')
3059
def _unescape_xml(data):
3060
"""Unescape predefined XML entities in a string of data."""
3062
if _unescape_re is None:
3063
_unescape_re = re.compile('\&([^;]*);')
3064
return _unescape_re.sub(_unescaper, data)
3067
class _VersionedFileChecker(object):
3069
def __init__(self, repository):
3070
self.repository = repository
3071
self.text_index = self.repository._generate_text_key_index()
3073
def calculate_file_version_parents(self, revision_id, file_id):
3074
"""Calculate the correct parents for a file version according to
3077
parent_keys = self.text_index[(file_id, revision_id)]
3078
if parent_keys == [_mod_revision.NULL_REVISION]:
3080
# strip the file_id, for the weave api
3081
return tuple([revision_id for file_id, revision_id in parent_keys])
3083
def check_file_version_parents(self, weave, file_id):
3084
"""Check the parents stored in a versioned file are correct.
3086
It also detects file versions that are not referenced by their
3087
corresponding revision's inventory.
3089
:returns: A tuple of (wrong_parents, dangling_file_versions).
3090
wrong_parents is a dict mapping {revision_id: (stored_parents,
3091
correct_parents)} for each revision_id where the stored parents
3092
are not correct. dangling_file_versions is a set of (file_id,
3093
revision_id) tuples for versions that are present in this versioned
3094
file, but not used by the corresponding inventory.
3097
unused_versions = set()
3098
versions = weave.versions()
3099
parent_map = weave.get_parent_map(versions)
3100
for num, revision_id in enumerate(versions):
3102
correct_parents = self.calculate_file_version_parents(
3103
revision_id, file_id)
3105
# The version is not part of the used keys.
3106
unused_versions.add(revision_id)
3109
knit_parents = tuple(parent_map[revision_id])
3110
except errors.RevisionNotPresent:
3112
if correct_parents != knit_parents:
3113
wrong_parents[revision_id] = (knit_parents, correct_parents)
3114
return wrong_parents, unused_versions
3117
def _old_get_graph(repository, revision_id):
3118
"""DO NOT USE. That is all. I'm serious."""
3119
graph = repository.get_graph()
3120
revision_graph = dict(((key, value) for key, value in
3121
graph.iter_ancestry([revision_id]) if value is not None))
3122
return _strip_NULL_ghosts(revision_graph)
3125
def _strip_NULL_ghosts(revision_graph):
3126
"""Also don't use this. more compatibility code for unmigrated clients."""
3127
# Filter ghosts, and null:
3128
if _mod_revision.NULL_REVISION in revision_graph:
3129
del revision_graph[_mod_revision.NULL_REVISION]
3130
for key, parents in revision_graph.items():
3131
revision_graph[key] = tuple(parent for parent in parents if parent
3133
return revision_graph
added
removed
bzrlib/repository.py
