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
revisions_inserted = 0
838
for item_key, bytes in stream:
839
if item_key[0] == 'file':
840
(file_id,) = item_key[1:]
841
knit = self.weave_store.get_weave_or_empty(
842
file_id, self.get_transaction())
843
elif item_key == ('inventory',):
844
knit = self.get_inventory_weave()
845
elif item_key == ('revisions',):
846
knit = self._revision_store.get_revision_file(
847
self.get_transaction())
848
revisions_inserted += 1
849
elif item_key == ('signatures',):
850
knit = self._revision_store.get_signature_file(
851
self.get_transaction())
853
raise errors.RepositoryDataStreamError(
854
"Unrecognised data stream key '%s'" % (item_key,))
855
decoded_list = bencode.bdecode(bytes)
856
format = decoded_list.pop(0)
859
for version, options, parents, some_bytes in decoded_list:
860
data_list.append((version, options, len(some_bytes), parents))
861
knit_bytes += some_bytes
862
buffer = StringIO(knit_bytes)
863
def reader_func(count):
867
return buffer.read(count)
868
knit.insert_data_stream(
869
(format, data_list, reader_func))
870
return revisions_inserted
873
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
874
"""Return the revision ids that other has that this does not.
876
These are returned in topological order.
878
revision_id: only return revision ids included by revision_id.
880
return InterRepository.get(other, self).search_missing_revision_ids(
881
revision_id, find_ghosts)
883
@deprecated_method(symbol_versioning.one_two)
885
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
886
"""Return the revision ids that other has that this does not.
888
These are returned in topological order.
890
revision_id: only return revision ids included by revision_id.
892
keys = self.search_missing_revision_ids(
893
other, revision_id, find_ghosts).get_keys()
896
parents = other.get_graph().get_parent_map(keys)
899
return tsort.topo_sort(parents)
903
"""Open the repository rooted at base.
905
For instance, if the repository is at URL/.bzr/repository,
906
Repository.open(URL) -> a Repository instance.
908
control = bzrdir.BzrDir.open(base)
909
return control.open_repository()
911
def copy_content_into(self, destination, revision_id=None):
912
"""Make a complete copy of the content in self into destination.
914
This is a destructive operation! Do not use it on existing
917
return InterRepository.get(self, destination).copy_content(revision_id)
919
def commit_write_group(self):
920
"""Commit the contents accrued within the current write group.
922
:seealso: start_write_group.
924
if self._write_group is not self.get_transaction():
925
# has an unlock or relock occured ?
926
raise errors.BzrError('mismatched lock context %r and '
928
(self.get_transaction(), self._write_group))
929
self._commit_write_group()
930
self._write_group = None
932
def _commit_write_group(self):
933
"""Template method for per-repository write group cleanup.
935
This is called before the write group is considered to be
936
finished and should ensure that all data handed to the repository
937
for writing during the write group is safely committed (to the
938
extent possible considering file system caching etc).
941
def fetch(self, source, revision_id=None, pb=None, find_ghosts=False):
942
"""Fetch the content required to construct revision_id from source.
944
If revision_id is None all content is copied.
945
:param find_ghosts: Find and copy revisions in the source that are
946
ghosts in the target (and not reachable directly by walking out to
947
the first-present revision in target from revision_id).
949
# fast path same-url fetch operations
950
if self.has_same_location(source):
951
# check that last_revision is in 'from' and then return a
953
if (revision_id is not None and
954
not _mod_revision.is_null(revision_id)):
955
self.get_revision(revision_id)
957
inter = InterRepository.get(source, self)
959
return inter.fetch(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts)
960
except NotImplementedError:
961
raise errors.IncompatibleRepositories(source, self)
963
def create_bundle(self, target, base, fileobj, format=None):
964
return serializer.write_bundle(self, target, base, fileobj, format)
966
def get_commit_builder(self, branch, parents, config, timestamp=None,
967
timezone=None, committer=None, revprops=None,
969
"""Obtain a CommitBuilder for this repository.
971
:param branch: Branch to commit to.
972
:param parents: Revision ids of the parents of the new revision.
973
:param config: Configuration to use.
974
:param timestamp: Optional timestamp recorded for commit.
975
:param timezone: Optional timezone for timestamp.
976
:param committer: Optional committer to set for commit.
977
:param revprops: Optional dictionary of revision properties.
978
:param revision_id: Optional revision id.
980
result = self._commit_builder_class(self, parents, config,
981
timestamp, timezone, committer, revprops, revision_id)
982
self.start_write_group()
986
if (self.control_files._lock_count == 1 and
987
self.control_files._lock_mode == 'w'):
988
if self._write_group is not None:
989
self.abort_write_group()
990
self.control_files.unlock()
991
raise errors.BzrError(
992
'Must end write groups before releasing write locks.')
993
self.control_files.unlock()
996
def clone(self, a_bzrdir, revision_id=None):
997
"""Clone this repository into a_bzrdir using the current format.
999
Currently no check is made that the format of this repository and
1000
the bzrdir format are compatible. FIXME RBC 20060201.
1002
:return: The newly created destination repository.
1004
# TODO: deprecate after 0.16; cloning this with all its settings is
1005
# probably not very useful -- mbp 20070423
1006
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
1007
self.copy_content_into(dest_repo, revision_id)
1010
def start_write_group(self):
1011
"""Start a write group in the repository.
1013
Write groups are used by repositories which do not have a 1:1 mapping
1014
between file ids and backend store to manage the insertion of data from
1015
both fetch and commit operations.
1017
A write lock is required around the start_write_group/commit_write_group
1018
for the support of lock-requiring repository formats.
1020
One can only insert data into a repository inside a write group.
1024
if not self.is_write_locked():
1025
raise errors.NotWriteLocked(self)
1026
if self._write_group:
1027
raise errors.BzrError('already in a write group')
1028
self._start_write_group()
1029
# so we can detect unlock/relock - the write group is now entered.
1030
self._write_group = self.get_transaction()
1032
def _start_write_group(self):
1033
"""Template method for per-repository write group startup.
1035
This is called before the write group is considered to be
1040
def sprout(self, to_bzrdir, revision_id=None):
1041
"""Create a descendent repository for new development.
1043
Unlike clone, this does not copy the settings of the repository.
1045
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
1046
dest_repo.fetch(self, revision_id=revision_id)
1049
def _create_sprouting_repo(self, a_bzrdir, shared):
1050
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
1051
# use target default format.
1052
dest_repo = a_bzrdir.create_repository()
1054
# Most control formats need the repository to be specifically
1055
# created, but on some old all-in-one formats it's not needed
1057
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
1058
except errors.UninitializableFormat:
1059
dest_repo = a_bzrdir.open_repository()
1063
def has_revision(self, revision_id):
1064
"""True if this repository has a copy of the revision."""
1065
return revision_id in self.has_revisions((revision_id,))
1067
def has_revisions(self, revision_ids):
1068
"""Probe to find out the presence of multiple revisions.
1070
:param revision_ids: An iterable of revision_ids.
1071
:return: A set of the revision_ids that were present.
1073
raise NotImplementedError(self.has_revisions)
1075
return self._revision_store.has_revision_id(revision_id,
1076
self.get_transaction())
1079
def get_revision(self, revision_id):
1080
"""Return the Revision object for a named revision."""
1081
return self.get_revisions([revision_id])[0]
1084
def get_revision_reconcile(self, revision_id):
1085
"""'reconcile' helper routine that allows access to a revision always.
1087
This variant of get_revision does not cross check the weave graph
1088
against the revision one as get_revision does: but it should only
1089
be used by reconcile, or reconcile-alike commands that are correcting
1090
or testing the revision graph.
1092
return self._get_revisions([revision_id])[0]
1095
def get_revisions(self, revision_ids):
1096
"""Get many revisions at once."""
1097
return self._get_revisions(revision_ids)
1100
def _get_revisions(self, revision_ids):
1101
"""Core work logic to get many revisions without sanity checks."""
1102
for rev_id in revision_ids:
1103
if not rev_id or not isinstance(rev_id, basestring):
1104
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1105
revs = self._revision_store.get_revisions(revision_ids,
1106
self.get_transaction())
1110
def get_revision_xml(self, revision_id):
1111
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1112
# would have already do it.
1113
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1114
rev = self.get_revision(revision_id)
1115
rev_tmp = StringIO()
1116
# the current serializer..
1117
self._revision_store._serializer.write_revision(rev, rev_tmp)
1119
return rev_tmp.getvalue()
1121
def get_deltas_for_revisions(self, revisions):
1122
"""Produce a generator of revision deltas.
1124
Note that the input is a sequence of REVISIONS, not revision_ids.
1125
Trees will be held in memory until the generator exits.
1126
Each delta is relative to the revision's lefthand predecessor.
1128
required_trees = set()
1129
for revision in revisions:
1130
required_trees.add(revision.revision_id)
1131
required_trees.update(revision.parent_ids[:1])
1132
trees = dict((t.get_revision_id(), t) for
1133
t in self.revision_trees(required_trees))
1134
for revision in revisions:
1135
if not revision.parent_ids:
1136
old_tree = self.revision_tree(None)
1138
old_tree = trees[revision.parent_ids[0]]
1139
yield trees[revision.revision_id].changes_from(old_tree)
1142
def get_revision_delta(self, revision_id):
1143
"""Return the delta for one revision.
1145
The delta is relative to the left-hand predecessor of the
1148
r = self.get_revision(revision_id)
1149
return list(self.get_deltas_for_revisions([r]))[0]
1152
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1153
signature = gpg_strategy.sign(plaintext)
1154
self.add_signature_text(revision_id, signature)
1157
def add_signature_text(self, revision_id, signature):
1158
self._revision_store.add_revision_signature_text(revision_id,
1160
self.get_transaction())
1162
def find_text_key_references(self):
1163
"""Find the text key references within the repository.
1165
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1166
revision_ids. Each altered file-ids has the exact revision_ids that
1167
altered it listed explicitly.
1168
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1169
to whether they were referred to by the inventory of the
1170
revision_id that they contain. The inventory texts from all present
1171
revision ids are assessed to generate this report.
1173
revision_ids = self.all_revision_ids()
1174
w = self.get_inventory_weave()
1175
pb = ui.ui_factory.nested_progress_bar()
1177
return self._find_text_key_references_from_xml_inventory_lines(
1178
w.iter_lines_added_or_present_in_versions(revision_ids, pb=pb))
1182
def _find_text_key_references_from_xml_inventory_lines(self,
1184
"""Core routine for extracting references to texts from inventories.
1186
This performs the translation of xml lines to revision ids.
1188
:param line_iterator: An iterator of lines, origin_version_id
1189
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1190
to whether they were referred to by the inventory of the
1191
revision_id that they contain. Note that if that revision_id was
1192
not part of the line_iterator's output then False will be given -
1193
even though it may actually refer to that key.
1195
if not self._serializer.support_altered_by_hack:
1196
raise AssertionError(
1197
"_find_text_key_references_from_xml_inventory_lines only "
1198
"supported for branches which store inventory as unnested xml"
1199
", not on %r" % self)
1202
# this code needs to read every new line in every inventory for the
1203
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1204
# not present in one of those inventories is unnecessary but not
1205
# harmful because we are filtering by the revision id marker in the
1206
# inventory lines : we only select file ids altered in one of those
1207
# revisions. We don't need to see all lines in the inventory because
1208
# only those added in an inventory in rev X can contain a revision=X
1210
unescape_revid_cache = {}
1211
unescape_fileid_cache = {}
1213
# jam 20061218 In a big fetch, this handles hundreds of thousands
1214
# of lines, so it has had a lot of inlining and optimizing done.
1215
# Sorry that it is a little bit messy.
1216
# Move several functions to be local variables, since this is a long
1218
search = self._file_ids_altered_regex.search
1219
unescape = _unescape_xml
1220
setdefault = result.setdefault
1221
for line, version_id in line_iterator:
1222
match = search(line)
1225
# One call to match.group() returning multiple items is quite a
1226
# bit faster than 2 calls to match.group() each returning 1
1227
file_id, revision_id = match.group('file_id', 'revision_id')
1229
# Inlining the cache lookups helps a lot when you make 170,000
1230
# lines and 350k ids, versus 8.4 unique ids.
1231
# Using a cache helps in 2 ways:
1232
# 1) Avoids unnecessary decoding calls
1233
# 2) Re-uses cached strings, which helps in future set and
1235
# (2) is enough that removing encoding entirely along with
1236
# the cache (so we are using plain strings) results in no
1237
# performance improvement.
1239
revision_id = unescape_revid_cache[revision_id]
1241
unescaped = unescape(revision_id)
1242
unescape_revid_cache[revision_id] = unescaped
1243
revision_id = unescaped
1245
# Note that unconditionally unescaping means that we deserialise
1246
# every fileid, which for general 'pull' is not great, but we don't
1247
# really want to have some many fulltexts that this matters anyway.
1250
file_id = unescape_fileid_cache[file_id]
1252
unescaped = unescape(file_id)
1253
unescape_fileid_cache[file_id] = unescaped
1256
key = (file_id, revision_id)
1257
setdefault(key, False)
1258
if revision_id == version_id:
1262
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1264
"""Helper routine for fileids_altered_by_revision_ids.
1266
This performs the translation of xml lines to revision ids.
1268
:param line_iterator: An iterator of lines, origin_version_id
1269
:param revision_ids: The revision ids to filter for. This should be a
1270
set or other type which supports efficient __contains__ lookups, as
1271
the revision id from each parsed line will be looked up in the
1272
revision_ids filter.
1273
:return: a dictionary mapping altered file-ids to an iterable of
1274
revision_ids. Each altered file-ids has the exact revision_ids that
1275
altered it listed explicitly.
1278
setdefault = result.setdefault
1279
for file_id, revision_id in \
1280
self._find_text_key_references_from_xml_inventory_lines(
1281
line_iterator).iterkeys():
1282
# once data is all ensured-consistent; then this is
1283
# if revision_id == version_id
1284
if revision_id in revision_ids:
1285
setdefault(file_id, set()).add(revision_id)
1288
def fileids_altered_by_revision_ids(self, revision_ids, _inv_weave=None):
1289
"""Find the file ids and versions affected by revisions.
1291
:param revisions: an iterable containing revision ids.
1292
:param _inv_weave: The inventory weave from this repository or None.
1293
If None, the inventory weave will be opened automatically.
1294
:return: a dictionary mapping altered file-ids to an iterable of
1295
revision_ids. Each altered file-ids has the exact revision_ids that
1296
altered it listed explicitly.
1298
selected_revision_ids = set(revision_ids)
1299
w = _inv_weave or self.get_inventory_weave()
1300
pb = ui.ui_factory.nested_progress_bar()
1302
return self._find_file_ids_from_xml_inventory_lines(
1303
w.iter_lines_added_or_present_in_versions(
1304
selected_revision_ids, pb=pb),
1305
selected_revision_ids)
1309
def iter_files_bytes(self, desired_files):
1310
"""Iterate through file versions.
1312
Files will not necessarily be returned in the order they occur in
1313
desired_files. No specific order is guaranteed.
1315
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1316
value supplied by the caller as part of desired_files. It should
1317
uniquely identify the file version in the caller's context. (Examples:
1318
an index number or a TreeTransform trans_id.)
1320
bytes_iterator is an iterable of bytestrings for the file. The
1321
kind of iterable and length of the bytestrings are unspecified, but for
1322
this implementation, it is a list of lines produced by
1323
VersionedFile.get_lines().
1325
:param desired_files: a list of (file_id, revision_id, identifier)
1328
transaction = self.get_transaction()
1329
for file_id, revision_id, callable_data in desired_files:
1331
weave = self.weave_store.get_weave(file_id, transaction)
1332
except errors.NoSuchFile:
1333
raise errors.NoSuchIdInRepository(self, file_id)
1334
yield callable_data, weave.get_lines(revision_id)
1336
def _generate_text_key_index(self, text_key_references=None,
1338
"""Generate a new text key index for the repository.
1340
This is an expensive function that will take considerable time to run.
1342
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1343
list of parents, also text keys. When a given key has no parents,
1344
the parents list will be [NULL_REVISION].
1346
# All revisions, to find inventory parents.
1347
if ancestors is None:
1348
graph = self.get_graph()
1349
ancestors = graph.get_parent_map(self.all_revision_ids())
1350
if text_key_references is None:
1351
text_key_references = self.find_text_key_references()
1352
pb = ui.ui_factory.nested_progress_bar()
1354
return self._do_generate_text_key_index(ancestors,
1355
text_key_references, pb)
1359
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1360
"""Helper for _generate_text_key_index to avoid deep nesting."""
1361
revision_order = tsort.topo_sort(ancestors)
1362
invalid_keys = set()
1364
for revision_id in revision_order:
1365
revision_keys[revision_id] = set()
1366
text_count = len(text_key_references)
1367
# a cache of the text keys to allow reuse; costs a dict of all the
1368
# keys, but saves a 2-tuple for every child of a given key.
1370
for text_key, valid in text_key_references.iteritems():
1372
invalid_keys.add(text_key)
1374
revision_keys[text_key[1]].add(text_key)
1375
text_key_cache[text_key] = text_key
1376
del text_key_references
1378
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1379
NULL_REVISION = _mod_revision.NULL_REVISION
1380
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1381
# too small for large or very branchy trees. However, for 55K path
1382
# trees, it would be easy to use too much memory trivially. Ideally we
1383
# could gauge this by looking at available real memory etc, but this is
1384
# always a tricky proposition.
1385
inventory_cache = lru_cache.LRUCache(10)
1386
batch_size = 10 # should be ~150MB on a 55K path tree
1387
batch_count = len(revision_order) / batch_size + 1
1389
pb.update("Calculating text parents.", processed_texts, text_count)
1390
for offset in xrange(batch_count):
1391
to_query = revision_order[offset * batch_size:(offset + 1) *
1395
for rev_tree in self.revision_trees(to_query):
1396
revision_id = rev_tree.get_revision_id()
1397
parent_ids = ancestors[revision_id]
1398
for text_key in revision_keys[revision_id]:
1399
pb.update("Calculating text parents.", processed_texts)
1400
processed_texts += 1
1401
candidate_parents = []
1402
for parent_id in parent_ids:
1403
parent_text_key = (text_key[0], parent_id)
1405
check_parent = parent_text_key not in \
1406
revision_keys[parent_id]
1408
# the parent parent_id is a ghost:
1409
check_parent = False
1410
# truncate the derived graph against this ghost.
1411
parent_text_key = None
1413
# look at the parent commit details inventories to
1414
# determine possible candidates in the per file graph.
1417
inv = inventory_cache[parent_id]
1419
inv = self.revision_tree(parent_id).inventory
1420
inventory_cache[parent_id] = inv
1421
parent_entry = inv._byid.get(text_key[0], None)
1422
if parent_entry is not None:
1424
text_key[0], parent_entry.revision)
1426
parent_text_key = None
1427
if parent_text_key is not None:
1428
candidate_parents.append(
1429
text_key_cache[parent_text_key])
1430
parent_heads = text_graph.heads(candidate_parents)
1431
new_parents = list(parent_heads)
1432
new_parents.sort(key=lambda x:candidate_parents.index(x))
1433
if new_parents == []:
1434
new_parents = [NULL_REVISION]
1435
text_index[text_key] = new_parents
1437
for text_key in invalid_keys:
1438
text_index[text_key] = [NULL_REVISION]
1441
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1442
"""Get an iterable listing the keys of all the data introduced by a set
1445
The keys will be ordered so that the corresponding items can be safely
1446
fetched and inserted in that order.
1448
:returns: An iterable producing tuples of (knit-kind, file-id,
1449
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1450
'revisions'. file-id is None unless knit-kind is 'file'.
1452
# XXX: it's a bit weird to control the inventory weave caching in this
1453
# generator. Ideally the caching would be done in fetch.py I think. Or
1454
# maybe this generator should explicitly have the contract that it
1455
# should not be iterated until the previously yielded item has been
1457
inv_w = self.get_inventory_weave()
1459
# file ids that changed
1460
file_ids = self.fileids_altered_by_revision_ids(revision_ids, inv_w)
1462
num_file_ids = len(file_ids)
1463
for file_id, altered_versions in file_ids.iteritems():
1464
if _files_pb is not None:
1465
_files_pb.update("fetch texts", count, num_file_ids)
1467
yield ("file", file_id, altered_versions)
1468
# We're done with the files_pb. Note that it finished by the caller,
1469
# just as it was created by the caller.
1473
yield ("inventory", None, revision_ids)
1476
revisions_with_signatures = set()
1477
for rev_id in revision_ids:
1479
self.get_signature_text(rev_id)
1480
except errors.NoSuchRevision:
1484
revisions_with_signatures.add(rev_id)
1485
yield ("signatures", None, revisions_with_signatures)
1488
yield ("revisions", None, revision_ids)
1491
def get_inventory_weave(self):
1492
return self.control_weaves.get_weave('inventory',
1493
self.get_transaction())
1496
def get_inventory(self, revision_id):
1497
"""Get Inventory object by revision id."""
1498
return self.iter_inventories([revision_id]).next()
1500
def iter_inventories(self, revision_ids):
1501
"""Get many inventories by revision_ids.
1503
This will buffer some or all of the texts used in constructing the
1504
inventories in memory, but will only parse a single inventory at a
1507
:return: An iterator of inventories.
1509
if ((None in revision_ids)
1510
or (_mod_revision.NULL_REVISION in revision_ids)):
1511
raise ValueError('cannot get null revision inventory')
1512
return self._iter_inventories(revision_ids)
1514
def _iter_inventories(self, revision_ids):
1515
"""single-document based inventory iteration."""
1516
texts = self.get_inventory_weave().get_texts(revision_ids)
1517
for text, revision_id in zip(texts, revision_ids):
1518
yield self.deserialise_inventory(revision_id, text)
1520
def deserialise_inventory(self, revision_id, xml):
1521
"""Transform the xml into an inventory object.
1523
:param revision_id: The expected revision id of the inventory.
1524
:param xml: A serialised inventory.
1526
result = self._serializer.read_inventory_from_string(xml, revision_id)
1527
if result.revision_id != revision_id:
1528
raise AssertionError('revision id mismatch %s != %s' % (
1529
result.revision_id, revision_id))
1532
def serialise_inventory(self, inv):
1533
return self._serializer.write_inventory_to_string(inv)
1535
def _serialise_inventory_to_lines(self, inv):
1536
return self._serializer.write_inventory_to_lines(inv)
1538
def get_serializer_format(self):
1539
return self._serializer.format_num
1542
def get_inventory_xml(self, revision_id):
1543
"""Get inventory XML as a file object."""
1545
iw = self.get_inventory_weave()
1546
return iw.get_text(revision_id)
1548
raise errors.HistoryMissing(self, 'inventory', revision_id)
1551
def get_inventory_sha1(self, revision_id):
1552
"""Return the sha1 hash of the inventory entry
1554
return self.get_revision(revision_id).inventory_sha1
1556
def iter_reverse_revision_history(self, revision_id):
1557
"""Iterate backwards through revision ids in the lefthand history
1559
:param revision_id: The revision id to start with. All its lefthand
1560
ancestors will be traversed.
1562
graph = self.get_graph()
1563
next_id = revision_id
1565
if next_id in (None, _mod_revision.NULL_REVISION):
1568
# Note: The following line may raise KeyError in the event of
1569
# truncated history. We decided not to have a try:except:raise
1570
# RevisionNotPresent here until we see a use for it, because of the
1571
# cost in an inner loop that is by its very nature O(history).
1572
# Robert Collins 20080326
1573
parents = graph.get_parent_map([next_id])[next_id]
1574
if len(parents) == 0:
1577
next_id = parents[0]
1580
def get_revision_inventory(self, revision_id):
1581
"""Return inventory of a past revision."""
1582
# TODO: Unify this with get_inventory()
1583
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1584
# must be the same as its revision, so this is trivial.
1585
if revision_id is None:
1586
# This does not make sense: if there is no revision,
1587
# then it is the current tree inventory surely ?!
1588
# and thus get_root_id() is something that looks at the last
1589
# commit on the branch, and the get_root_id is an inventory check.
1590
raise NotImplementedError
1591
# return Inventory(self.get_root_id())
1593
return self.get_inventory(revision_id)
1596
def is_shared(self):
1597
"""Return True if this repository is flagged as a shared repository."""
1598
raise NotImplementedError(self.is_shared)
1601
def reconcile(self, other=None, thorough=False):
1602
"""Reconcile this repository."""
1603
from bzrlib.reconcile import RepoReconciler
1604
reconciler = RepoReconciler(self, thorough=thorough)
1605
reconciler.reconcile()
1608
def _refresh_data(self):
1609
"""Helper called from lock_* to ensure coherency with disk.
1611
The default implementation does nothing; it is however possible
1612
for repositories to maintain loaded indices across multiple locks
1613
by checking inside their implementation of this method to see
1614
whether their indices are still valid. This depends of course on
1615
the disk format being validatable in this manner.
1619
def revision_tree(self, revision_id):
1620
"""Return Tree for a revision on this branch.
1622
`revision_id` may be None for the empty tree revision.
1624
# TODO: refactor this to use an existing revision object
1625
# so we don't need to read it in twice.
1626
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1627
return RevisionTree(self, Inventory(root_id=None),
1628
_mod_revision.NULL_REVISION)
1630
inv = self.get_revision_inventory(revision_id)
1631
return RevisionTree(self, inv, revision_id)
1633
def revision_trees(self, revision_ids):
1634
"""Return Tree for a revision on this branch.
1636
`revision_id` may not be None or 'null:'"""
1637
inventories = self.iter_inventories(revision_ids)
1638
for inv in inventories:
1639
yield RevisionTree(self, inv, inv.revision_id)
1642
def get_ancestry(self, revision_id, topo_sorted=True):
1643
"""Return a list of revision-ids integrated by a revision.
1645
The first element of the list is always None, indicating the origin
1646
revision. This might change when we have history horizons, or
1647
perhaps we should have a new API.
1649
This is topologically sorted.
1651
if _mod_revision.is_null(revision_id):
1653
if not self.has_revision(revision_id):
1654
raise errors.NoSuchRevision(self, revision_id)
1655
w = self.get_inventory_weave()
1656
candidates = w.get_ancestry(revision_id, topo_sorted)
1657
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1660
"""Compress the data within the repository.
1662
This operation only makes sense for some repository types. For other
1663
types it should be a no-op that just returns.
1665
This stub method does not require a lock, but subclasses should use
1666
@needs_write_lock as this is a long running call its reasonable to
1667
implicitly lock for the user.
1671
def print_file(self, file, revision_id):
1672
"""Print `file` to stdout.
1674
FIXME RBC 20060125 as John Meinel points out this is a bad api
1675
- it writes to stdout, it assumes that that is valid etc. Fix
1676
by creating a new more flexible convenience function.
1678
tree = self.revision_tree(revision_id)
1679
# use inventory as it was in that revision
1680
file_id = tree.inventory.path2id(file)
1682
# TODO: jam 20060427 Write a test for this code path
1683
# it had a bug in it, and was raising the wrong
1685
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1686
tree.print_file(file_id)
1688
def get_transaction(self):
1689
return self.control_files.get_transaction()
1691
@deprecated_method(symbol_versioning.one_one)
1692
def get_parents(self, revision_ids):
1693
"""See StackedParentsProvider.get_parents"""
1694
parent_map = self.get_parent_map(revision_ids)
1695
return [parent_map.get(r, None) for r in revision_ids]
1697
def get_parent_map(self, keys):
1698
"""See graph._StackedParentsProvider.get_parent_map"""
1700
for revision_id in keys:
1701
if revision_id is None:
1702
raise ValueError('get_parent_map(None) is not valid')
1703
if revision_id == _mod_revision.NULL_REVISION:
1704
parent_map[revision_id] = ()
1707
parent_id_list = self.get_revision(revision_id).parent_ids
1708
except errors.NoSuchRevision:
1711
if len(parent_id_list) == 0:
1712
parent_ids = (_mod_revision.NULL_REVISION,)
1714
parent_ids = tuple(parent_id_list)
1715
parent_map[revision_id] = parent_ids
1718
def _make_parents_provider(self):
1721
def get_graph(self, other_repository=None):
1722
"""Return the graph walker for this repository format"""
1723
parents_provider = self._make_parents_provider()
1724
if (other_repository is not None and
1725
not self.has_same_location(other_repository)):
1726
parents_provider = graph._StackedParentsProvider(
1727
[parents_provider, other_repository._make_parents_provider()])
1728
return graph.Graph(parents_provider)
1730
def _get_versioned_file_checker(self):
1731
"""Return an object suitable for checking versioned files."""
1732
return _VersionedFileChecker(self)
1734
def revision_ids_to_search_result(self, result_set):
1735
"""Convert a set of revision ids to a graph SearchResult."""
1736
result_parents = set()
1737
for parents in self.get_graph().get_parent_map(
1738
result_set).itervalues():
1739
result_parents.update(parents)
1740
included_keys = result_set.intersection(result_parents)
1741
start_keys = result_set.difference(included_keys)
1742
exclude_keys = result_parents.difference(result_set)
1743
result = graph.SearchResult(start_keys, exclude_keys,
1744
len(result_set), result_set)
1748
def set_make_working_trees(self, new_value):
1749
"""Set the policy flag for making working trees when creating branches.
1751
This only applies to branches that use this repository.
1753
The default is 'True'.
1754
:param new_value: True to restore the default, False to disable making
1757
raise NotImplementedError(self.set_make_working_trees)
1759
def make_working_trees(self):
1760
"""Returns the policy for making working trees on new branches."""
1761
raise NotImplementedError(self.make_working_trees)
1764
def sign_revision(self, revision_id, gpg_strategy):
1765
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1766
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1769
def has_signature_for_revision_id(self, revision_id):
1770
"""Query for a revision signature for revision_id in the repository."""
1771
return self._revision_store.has_signature(revision_id,
1772
self.get_transaction())
1775
def get_signature_text(self, revision_id):
1776
"""Return the text for a signature."""
1777
return self._revision_store.get_signature_text(revision_id,
1778
self.get_transaction())
1781
def check(self, revision_ids=None):
1782
"""Check consistency of all history of given revision_ids.
1784
Different repository implementations should override _check().
1786
:param revision_ids: A non-empty list of revision_ids whose ancestry
1787
will be checked. Typically the last revision_id of a branch.
1789
return self._check(revision_ids)
1791
def _check(self, revision_ids):
1792
result = check.Check(self)
1796
def _warn_if_deprecated(self):
1797
global _deprecation_warning_done
1798
if _deprecation_warning_done:
1800
_deprecation_warning_done = True
1801
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1802
% (self._format, self.bzrdir.transport.base))
1804
def supports_rich_root(self):
1805
return self._format.rich_root_data
1807
def _check_ascii_revisionid(self, revision_id, method):
1808
"""Private helper for ascii-only repositories."""
1809
# weave repositories refuse to store revisionids that are non-ascii.
1810
if revision_id is not None:
1811
# weaves require ascii revision ids.
1812
if isinstance(revision_id, unicode):
1814
revision_id.encode('ascii')
1815
except UnicodeEncodeError:
1816
raise errors.NonAsciiRevisionId(method, self)
1819
revision_id.decode('ascii')
1820
except UnicodeDecodeError:
1821
raise errors.NonAsciiRevisionId(method, self)
1823
def revision_graph_can_have_wrong_parents(self):
1824
"""Is it possible for this repository to have a revision graph with
1827
If True, then this repository must also implement
1828
_find_inconsistent_revision_parents so that check and reconcile can
1829
check for inconsistencies before proceeding with other checks that may
1830
depend on the revision index being consistent.
1832
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1835
# remove these delegates a while after bzr 0.15
1836
def __make_delegated(name, from_module):
1837
def _deprecated_repository_forwarder():
1838
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1839
% (name, from_module),
1842
m = __import__(from_module, globals(), locals(), [name])
1844
return getattr(m, name)
1845
except AttributeError:
1846
raise AttributeError('module %s has no name %s'
1848
globals()[name] = _deprecated_repository_forwarder
1851
'AllInOneRepository',
1852
'WeaveMetaDirRepository',
1853
'PreSplitOutRepositoryFormat',
1854
'RepositoryFormat4',
1855
'RepositoryFormat5',
1856
'RepositoryFormat6',
1857
'RepositoryFormat7',
1859
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1863
'RepositoryFormatKnit',
1864
'RepositoryFormatKnit1',
1866
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1869
def install_revision(repository, rev, revision_tree):
1870
"""Install all revision data into a repository."""
1871
install_revisions(repository, [(rev, revision_tree, None)])
1874
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1875
"""Install all revision data into a repository.
1877
Accepts an iterable of revision, tree, signature tuples. The signature
1880
repository.start_write_group()
1882
for n, (revision, revision_tree, signature) in enumerate(iterable):
1883
_install_revision(repository, revision, revision_tree, signature)
1885
pb.update('Transferring revisions', n + 1, num_revisions)
1887
repository.abort_write_group()
1890
repository.commit_write_group()
1893
def _install_revision(repository, rev, revision_tree, signature):
1894
"""Install all revision data into a repository."""
1895
present_parents = []
1897
for p_id in rev.parent_ids:
1898
if repository.has_revision(p_id):
1899
present_parents.append(p_id)
1900
parent_trees[p_id] = repository.revision_tree(p_id)
1902
parent_trees[p_id] = repository.revision_tree(None)
1904
inv = revision_tree.inventory
1905
entries = inv.iter_entries()
1906
# backwards compatibility hack: skip the root id.
1907
if not repository.supports_rich_root():
1908
path, root = entries.next()
1909
if root.revision != rev.revision_id:
1910
raise errors.IncompatibleRevision(repr(repository))
1911
# Add the texts that are not already present
1912
for path, ie in entries:
1913
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1914
repository.get_transaction())
1915
if ie.revision not in w:
1917
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1918
# with InventoryEntry.find_previous_heads(). if it is, then there
1919
# is a latent bug here where the parents may have ancestors of each
1921
for revision, tree in parent_trees.iteritems():
1922
if ie.file_id not in tree:
1924
parent_id = tree.inventory[ie.file_id].revision
1925
if parent_id in text_parents:
1927
text_parents.append(parent_id)
1929
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1930
repository.get_transaction())
1931
lines = revision_tree.get_file(ie.file_id).readlines()
1932
vfile.add_lines(rev.revision_id, text_parents, lines)
1934
# install the inventory
1935
repository.add_inventory(rev.revision_id, inv, present_parents)
1936
except errors.RevisionAlreadyPresent:
1938
if signature is not None:
1939
repository.add_signature_text(rev.revision_id, signature)
1940
repository.add_revision(rev.revision_id, rev, inv)
1943
class MetaDirRepository(Repository):
1944
"""Repositories in the new meta-dir layout.
1946
:ivar _transport: Transport for access to repository control files,
1947
typically pointing to .bzr/repository.
1950
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1951
super(MetaDirRepository, self).__init__(_format,
1957
self._transport = control_files._transport
1960
def is_shared(self):
1961
"""Return True if this repository is flagged as a shared repository."""
1962
return self._transport.has('shared-storage')
1965
def set_make_working_trees(self, new_value):
1966
"""Set the policy flag for making working trees when creating branches.
1968
This only applies to branches that use this repository.
1970
The default is 'True'.
1971
:param new_value: True to restore the default, False to disable making
1976
self._transport.delete('no-working-trees')
1977
except errors.NoSuchFile:
1980
self._transport.put_bytes('no-working-trees', '',
1981
mode=self.bzrdir._get_file_mode())
1983
def make_working_trees(self):
1984
"""Returns the policy for making working trees on new branches."""
1985
return not self._transport.has('no-working-trees')
1988
class MetaDirVersionedFileRepository(MetaDirRepository):
1989
"""Repositories in a meta-dir, that work via versioned file objects."""
1991
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1992
super(MetaDirVersionedFileRepository, self).__init__(_format, a_bzrdir,
1993
control_files, _revision_store, control_store, text_store)
1994
_revision_store.get_scope = self.get_transaction
1995
control_store.get_scope = self.get_transaction
1996
text_store.get_scope = self.get_transaction
1999
class RepositoryFormatRegistry(registry.Registry):
2000
"""Registry of RepositoryFormats."""
2002
def get(self, format_string):
2003
r = registry.Registry.get(self, format_string)
2009
format_registry = RepositoryFormatRegistry()
2010
"""Registry of formats, indexed by their identifying format string.
2012
This can contain either format instances themselves, or classes/factories that
2013
can be called to obtain one.
2017
#####################################################################
2018
# Repository Formats
2020
class RepositoryFormat(object):
2021
"""A repository format.
2023
Formats provide three things:
2024
* An initialization routine to construct repository data on disk.
2025
* a format string which is used when the BzrDir supports versioned
2027
* an open routine which returns a Repository instance.
2029
There is one and only one Format subclass for each on-disk format. But
2030
there can be one Repository subclass that is used for several different
2031
formats. The _format attribute on a Repository instance can be used to
2032
determine the disk format.
2034
Formats are placed in an dict by their format string for reference
2035
during opening. These should be subclasses of RepositoryFormat
2038
Once a format is deprecated, just deprecate the initialize and open
2039
methods on the format class. Do not deprecate the object, as the
2040
object will be created every system load.
2042
Common instance attributes:
2043
_matchingbzrdir - the bzrdir format that the repository format was
2044
originally written to work with. This can be used if manually
2045
constructing a bzrdir and repository, or more commonly for test suite
2049
# Set to True or False in derived classes. True indicates that the format
2050
# supports ghosts gracefully.
2051
supports_ghosts = None
2052
# Can this repository be given external locations to lookup additional
2053
# data. Set to True or False in derived classes.
2054
supports_external_lookups = None
2057
return "<%s>" % self.__class__.__name__
2059
def __eq__(self, other):
2060
# format objects are generally stateless
2061
return isinstance(other, self.__class__)
2063
def __ne__(self, other):
2064
return not self == other
2067
def find_format(klass, a_bzrdir):
2068
"""Return the format for the repository object in a_bzrdir.
2070
This is used by bzr native formats that have a "format" file in
2071
the repository. Other methods may be used by different types of
2075
transport = a_bzrdir.get_repository_transport(None)
2076
format_string = transport.get("format").read()
2077
return format_registry.get(format_string)
2078
except errors.NoSuchFile:
2079
raise errors.NoRepositoryPresent(a_bzrdir)
2081
raise errors.UnknownFormatError(format=format_string,
2085
def register_format(klass, format):
2086
format_registry.register(format.get_format_string(), format)
2089
def unregister_format(klass, format):
2090
format_registry.remove(format.get_format_string())
2093
def get_default_format(klass):
2094
"""Return the current default format."""
2095
from bzrlib import bzrdir
2096
return bzrdir.format_registry.make_bzrdir('default').repository_format
2098
def _get_control_store(self, repo_transport, control_files):
2099
"""Return the control store for this repository."""
2100
raise NotImplementedError(self._get_control_store)
2102
def get_format_string(self):
2103
"""Return the ASCII format string that identifies this format.
2105
Note that in pre format ?? repositories the format string is
2106
not permitted nor written to disk.
2108
raise NotImplementedError(self.get_format_string)
2110
def get_format_description(self):
2111
"""Return the short description for this format."""
2112
raise NotImplementedError(self.get_format_description)
2114
def _get_revision_store(self, repo_transport, control_files):
2115
"""Return the revision store object for this a_bzrdir."""
2116
raise NotImplementedError(self._get_revision_store)
2118
def _get_text_rev_store(self,
2125
"""Common logic for getting a revision store for a repository.
2127
see self._get_revision_store for the subclass-overridable method to
2128
get the store for a repository.
2130
from bzrlib.store.revision.text import TextRevisionStore
2131
dir_mode = control_files._dir_mode
2132
file_mode = control_files._file_mode
2133
text_store = TextStore(transport.clone(name),
2135
compressed=compressed,
2137
file_mode=file_mode)
2138
_revision_store = TextRevisionStore(text_store, serializer)
2139
return _revision_store
2141
# TODO: this shouldn't be in the base class, it's specific to things that
2142
# use weaves or knits -- mbp 20070207
2143
def _get_versioned_file_store(self,
2148
versionedfile_class=None,
2149
versionedfile_kwargs={},
2151
if versionedfile_class is None:
2152
versionedfile_class = self._versionedfile_class
2153
weave_transport = control_files._transport.clone(name)
2154
dir_mode = control_files._dir_mode
2155
file_mode = control_files._file_mode
2156
return VersionedFileStore(weave_transport, prefixed=prefixed,
2158
file_mode=file_mode,
2159
versionedfile_class=versionedfile_class,
2160
versionedfile_kwargs=versionedfile_kwargs,
2163
def initialize(self, a_bzrdir, shared=False):
2164
"""Initialize a repository of this format in a_bzrdir.
2166
:param a_bzrdir: The bzrdir to put the new repository in it.
2167
:param shared: The repository should be initialized as a sharable one.
2168
:returns: The new repository object.
2170
This may raise UninitializableFormat if shared repository are not
2171
compatible the a_bzrdir.
2173
raise NotImplementedError(self.initialize)
2175
def is_supported(self):
2176
"""Is this format supported?
2178
Supported formats must be initializable and openable.
2179
Unsupported formats may not support initialization or committing or
2180
some other features depending on the reason for not being supported.
2184
def check_conversion_target(self, target_format):
2185
raise NotImplementedError(self.check_conversion_target)
2187
def open(self, a_bzrdir, _found=False):
2188
"""Return an instance of this format for the bzrdir a_bzrdir.
2190
_found is a private parameter, do not use it.
2192
raise NotImplementedError(self.open)
2195
class MetaDirRepositoryFormat(RepositoryFormat):
2196
"""Common base class for the new repositories using the metadir layout."""
2198
rich_root_data = False
2199
supports_tree_reference = False
2200
supports_external_lookups = False
2201
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
2204
super(MetaDirRepositoryFormat, self).__init__()
2206
def _create_control_files(self, a_bzrdir):
2207
"""Create the required files and the initial control_files object."""
2208
# FIXME: RBC 20060125 don't peek under the covers
2209
# NB: no need to escape relative paths that are url safe.
2210
repository_transport = a_bzrdir.get_repository_transport(self)
2211
control_files = lockable_files.LockableFiles(repository_transport,
2212
'lock', lockdir.LockDir)
2213
control_files.create_lock()
2214
return control_files
2216
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
2217
"""Upload the initial blank content."""
2218
control_files = self._create_control_files(a_bzrdir)
2219
control_files.lock_write()
2220
transport = control_files._transport
2222
utf8_files += [('shared-storage', '')]
2224
transport.mkdir_multi(dirs, mode=a_bzrdir._get_dir_mode())
2225
for (filename, content_stream) in files:
2226
transport.put_file(filename, content_stream,
2227
mode=a_bzrdir._get_file_mode())
2228
for (filename, content_bytes) in utf8_files:
2229
transport.put_bytes_non_atomic(filename, content_bytes,
2230
mode=a_bzrdir._get_file_mode())
2232
control_files.unlock()
2235
# formats which have no format string are not discoverable
2236
# and not independently creatable, so are not registered. They're
2237
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
2238
# needed, it's constructed directly by the BzrDir. Non-native formats where
2239
# the repository is not separately opened are similar.
2241
format_registry.register_lazy(
2242
'Bazaar-NG Repository format 7',
2243
'bzrlib.repofmt.weaverepo',
2247
format_registry.register_lazy(
2248
'Bazaar-NG Knit Repository Format 1',
2249
'bzrlib.repofmt.knitrepo',
2250
'RepositoryFormatKnit1',
2253
format_registry.register_lazy(
2254
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
2255
'bzrlib.repofmt.knitrepo',
2256
'RepositoryFormatKnit3',
2259
format_registry.register_lazy(
2260
'Bazaar Knit Repository Format 4 (bzr 1.0)\n',
2261
'bzrlib.repofmt.knitrepo',
2262
'RepositoryFormatKnit4',
2265
# Pack-based formats. There is one format for pre-subtrees, and one for
2266
# post-subtrees to allow ease of testing.
2267
# NOTE: These are experimental in 0.92. Stable in 1.0 and above
2268
format_registry.register_lazy(
2269
'Bazaar pack repository format 1 (needs bzr 0.92)\n',
2270
'bzrlib.repofmt.pack_repo',
2271
'RepositoryFormatKnitPack1',
2273
format_registry.register_lazy(
2274
'Bazaar pack repository format 1 with subtree support (needs bzr 0.92)\n',
2275
'bzrlib.repofmt.pack_repo',
2276
'RepositoryFormatKnitPack3',
2278
format_registry.register_lazy(
2279
'Bazaar pack repository format 1 with rich root (needs bzr 1.0)\n',
2280
'bzrlib.repofmt.pack_repo',
2281
'RepositoryFormatKnitPack4',
2283
# Development formats.
2285
# development 0 - stub to introduce development versioning scheme.
2286
format_registry.register_lazy(
2287
"Bazaar development format 0 (needs bzr.dev from before 1.3)\n",
2288
'bzrlib.repofmt.pack_repo',
2289
'RepositoryFormatPackDevelopment0',
2291
format_registry.register_lazy(
2292
("Bazaar development format 0 with subtree support "
2293
"(needs bzr.dev from before 1.3)\n"),
2294
'bzrlib.repofmt.pack_repo',
2295
'RepositoryFormatPackDevelopment0Subtree',
2297
# 1.3->1.4 go below here
2300
class InterRepository(InterObject):
2301
"""This class represents operations taking place between two repositories.
2303
Its instances have methods like copy_content and fetch, and contain
2304
references to the source and target repositories these operations can be
2307
Often we will provide convenience methods on 'repository' which carry out
2308
operations with another repository - they will always forward to
2309
InterRepository.get(other).method_name(parameters).
2313
"""The available optimised InterRepository types."""
2315
def copy_content(self, revision_id=None):
2316
raise NotImplementedError(self.copy_content)
2318
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2319
"""Fetch the content required to construct revision_id.
2321
The content is copied from self.source to self.target.
2323
:param revision_id: if None all content is copied, if NULL_REVISION no
2325
:param pb: optional progress bar to use for progress reports. If not
2326
provided a default one will be created.
2328
Returns the copied revision count and the failed revisions in a tuple:
2331
raise NotImplementedError(self.fetch)
2333
def _walk_to_common_revisions(self, revision_ids):
2334
"""Walk out from revision_ids in source to revisions target has.
2336
:param revision_ids: The start point for the search.
2337
:return: A set of revision ids.
2339
target_graph = self.target.get_graph()
2340
revision_ids = frozenset(revision_ids)
2341
if set(target_graph.get_parent_map(revision_ids)) == revision_ids:
2342
return graph.SearchResult(revision_ids, set(), 0, set())
2343
missing_revs = set()
2344
source_graph = self.source.get_graph()
2345
# ensure we don't pay silly lookup costs.
2346
searcher = source_graph._make_breadth_first_searcher(revision_ids)
2347
null_set = frozenset([_mod_revision.NULL_REVISION])
2350
next_revs, ghosts = searcher.next_with_ghosts()
2351
except StopIteration:
2353
if revision_ids.intersection(ghosts):
2354
absent_ids = set(revision_ids.intersection(ghosts))
2355
# If all absent_ids are present in target, no error is needed.
2356
absent_ids.difference_update(
2357
set(target_graph.get_parent_map(absent_ids)))
2359
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2360
# we don't care about other ghosts as we can't fetch them and
2361
# haven't been asked to.
2362
next_revs = set(next_revs)
2363
# we always have NULL_REVISION present.
2364
have_revs = set(target_graph.get_parent_map(next_revs)).union(null_set)
2365
missing_revs.update(next_revs - have_revs)
2366
searcher.stop_searching_any(have_revs)
2367
return searcher.get_result()
2369
@deprecated_method(symbol_versioning.one_two)
2371
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2372
"""Return the revision ids that source has that target does not.
2374
These are returned in topological order.
2376
:param revision_id: only return revision ids included by this
2378
:param find_ghosts: If True find missing revisions in deep history
2379
rather than just finding the surface difference.
2381
return list(self.search_missing_revision_ids(
2382
revision_id, find_ghosts).get_keys())
2385
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2386
"""Return the revision ids that source has that target does not.
2388
:param revision_id: only return revision ids included by this
2390
:param find_ghosts: If True find missing revisions in deep history
2391
rather than just finding the surface difference.
2392
:return: A bzrlib.graph.SearchResult.
2394
# stop searching at found target revisions.
2395
if not find_ghosts and revision_id is not None:
2396
return self._walk_to_common_revisions([revision_id])
2397
# generic, possibly worst case, slow code path.
2398
target_ids = set(self.target.all_revision_ids())
2399
if revision_id is not None:
2400
source_ids = self.source.get_ancestry(revision_id)
2401
if source_ids[0] is not None:
2402
raise AssertionError()
2405
source_ids = self.source.all_revision_ids()
2406
result_set = set(source_ids).difference(target_ids)
2407
return self.source.revision_ids_to_search_result(result_set)
2410
def _same_model(source, target):
2411
"""True if source and target have the same data representation."""
2412
if source.supports_rich_root() != target.supports_rich_root():
2414
if source._serializer != target._serializer:
2419
class InterSameDataRepository(InterRepository):
2420
"""Code for converting between repositories that represent the same data.
2422
Data format and model must match for this to work.
2426
def _get_repo_format_to_test(self):
2427
"""Repository format for testing with.
2429
InterSameData can pull from subtree to subtree and from non-subtree to
2430
non-subtree, so we test this with the richest repository format.
2432
from bzrlib.repofmt import knitrepo
2433
return knitrepo.RepositoryFormatKnit3()
2436
def is_compatible(source, target):
2437
return InterRepository._same_model(source, target)
2440
def copy_content(self, revision_id=None):
2441
"""Make a complete copy of the content in self into destination.
2443
This copies both the repository's revision data, and configuration information
2444
such as the make_working_trees setting.
2446
This is a destructive operation! Do not use it on existing
2449
:param revision_id: Only copy the content needed to construct
2450
revision_id and its parents.
2453
self.target.set_make_working_trees(self.source.make_working_trees())
2454
except NotImplementedError:
2456
# but don't bother fetching if we have the needed data now.
2457
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2458
self.target.has_revision(revision_id)):
2460
self.target.fetch(self.source, revision_id=revision_id)
2463
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2464
"""See InterRepository.fetch()."""
2465
from bzrlib.fetch import GenericRepoFetcher
2466
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2467
self.source, self.source._format, self.target,
2468
self.target._format)
2469
f = GenericRepoFetcher(to_repository=self.target,
2470
from_repository=self.source,
2471
last_revision=revision_id,
2472
pb=pb, find_ghosts=find_ghosts)
2473
return f.count_copied, f.failed_revisions
2476
class InterWeaveRepo(InterSameDataRepository):
2477
"""Optimised code paths between Weave based repositories.
2479
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2480
implemented lazy inter-object optimisation.
2484
def _get_repo_format_to_test(self):
2485
from bzrlib.repofmt import weaverepo
2486
return weaverepo.RepositoryFormat7()
2489
def is_compatible(source, target):
2490
"""Be compatible with known Weave formats.
2492
We don't test for the stores being of specific types because that
2493
could lead to confusing results, and there is no need to be
2496
from bzrlib.repofmt.weaverepo import (
2502
return (isinstance(source._format, (RepositoryFormat5,
2504
RepositoryFormat7)) and
2505
isinstance(target._format, (RepositoryFormat5,
2507
RepositoryFormat7)))
2508
except AttributeError:
2512
def copy_content(self, revision_id=None):
2513
"""See InterRepository.copy_content()."""
2514
# weave specific optimised path:
2516
self.target.set_make_working_trees(self.source.make_working_trees())
2517
except (errors.RepositoryUpgradeRequired, NotImplemented):
2519
# FIXME do not peek!
2520
if self.source._transport.listable():
2521
pb = ui.ui_factory.nested_progress_bar()
2523
self.target.weave_store.copy_all_ids(
2524
self.source.weave_store,
2526
from_transaction=self.source.get_transaction(),
2527
to_transaction=self.target.get_transaction())
2528
pb.update('copying inventory', 0, 1)
2529
self.target.control_weaves.copy_multi(
2530
self.source.control_weaves, ['inventory'],
2531
from_transaction=self.source.get_transaction(),
2532
to_transaction=self.target.get_transaction())
2533
self.target._revision_store.text_store.copy_all_ids(
2534
self.source._revision_store.text_store,
2539
self.target.fetch(self.source, revision_id=revision_id)
2542
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2543
"""See InterRepository.fetch()."""
2544
from bzrlib.fetch import GenericRepoFetcher
2545
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2546
self.source, self.source._format, self.target, self.target._format)
2547
f = GenericRepoFetcher(to_repository=self.target,
2548
from_repository=self.source,
2549
last_revision=revision_id,
2550
pb=pb, find_ghosts=find_ghosts)
2551
return f.count_copied, f.failed_revisions
2554
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2555
"""See InterRepository.missing_revision_ids()."""
2556
# we want all revisions to satisfy revision_id in source.
2557
# but we don't want to stat every file here and there.
2558
# we want then, all revisions other needs to satisfy revision_id
2559
# checked, but not those that we have locally.
2560
# so the first thing is to get a subset of the revisions to
2561
# satisfy revision_id in source, and then eliminate those that
2562
# we do already have.
2563
# this is slow on high latency connection to self, but as as this
2564
# disk format scales terribly for push anyway due to rewriting
2565
# inventory.weave, this is considered acceptable.
2567
if revision_id is not None:
2568
source_ids = self.source.get_ancestry(revision_id)
2569
if source_ids[0] is not None:
2570
raise AssertionError()
2573
source_ids = self.source._all_possible_ids()
2574
source_ids_set = set(source_ids)
2575
# source_ids is the worst possible case we may need to pull.
2576
# now we want to filter source_ids against what we actually
2577
# have in target, but don't try to check for existence where we know
2578
# we do not have a revision as that would be pointless.
2579
target_ids = set(self.target._all_possible_ids())
2580
possibly_present_revisions = target_ids.intersection(source_ids_set)
2581
actually_present_revisions = set(
2582
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2583
required_revisions = source_ids_set.difference(actually_present_revisions)
2584
if revision_id is not None:
2585
# we used get_ancestry to determine source_ids then we are assured all
2586
# revisions referenced are present as they are installed in topological order.
2587
# and the tip revision was validated by get_ancestry.
2588
result_set = required_revisions
2590
# if we just grabbed the possibly available ids, then
2591
# we only have an estimate of whats available and need to validate
2592
# that against the revision records.
2594
self.source._eliminate_revisions_not_present(required_revisions))
2595
return self.source.revision_ids_to_search_result(result_set)
2598
class InterKnitRepo(InterSameDataRepository):
2599
"""Optimised code paths between Knit based repositories."""
2602
def _get_repo_format_to_test(self):
2603
from bzrlib.repofmt import knitrepo
2604
return knitrepo.RepositoryFormatKnit1()
2607
def is_compatible(source, target):
2608
"""Be compatible with known Knit formats.
2610
We don't test for the stores being of specific types because that
2611
could lead to confusing results, and there is no need to be
2614
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2616
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2617
isinstance(target._format, RepositoryFormatKnit))
2618
except AttributeError:
2620
return are_knits and InterRepository._same_model(source, target)
2623
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2624
"""See InterRepository.fetch()."""
2625
from bzrlib.fetch import KnitRepoFetcher
2626
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2627
self.source, self.source._format, self.target, self.target._format)
2628
f = KnitRepoFetcher(to_repository=self.target,
2629
from_repository=self.source,
2630
last_revision=revision_id,
2631
pb=pb, find_ghosts=find_ghosts)
2632
return f.count_copied, f.failed_revisions
2635
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2636
"""See InterRepository.missing_revision_ids()."""
2637
if revision_id is not None:
2638
source_ids = self.source.get_ancestry(revision_id)
2639
if source_ids[0] is not None:
2640
raise AssertionError()
2643
source_ids = self.source.all_revision_ids()
2644
source_ids_set = set(source_ids)
2645
# source_ids is the worst possible case we may need to pull.
2646
# now we want to filter source_ids against what we actually
2647
# have in target, but don't try to check for existence where we know
2648
# we do not have a revision as that would be pointless.
2649
target_ids = set(self.target.all_revision_ids())
2650
possibly_present_revisions = target_ids.intersection(source_ids_set)
2651
actually_present_revisions = set(
2652
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2653
required_revisions = source_ids_set.difference(actually_present_revisions)
2654
if revision_id is not None:
2655
# we used get_ancestry to determine source_ids then we are assured all
2656
# revisions referenced are present as they are installed in topological order.
2657
# and the tip revision was validated by get_ancestry.
2658
result_set = required_revisions
2660
# if we just grabbed the possibly available ids, then
2661
# we only have an estimate of whats available and need to validate
2662
# that against the revision records.
2664
self.source._eliminate_revisions_not_present(required_revisions))
2665
return self.source.revision_ids_to_search_result(result_set)
2668
class InterPackRepo(InterSameDataRepository):
2669
"""Optimised code paths between Pack based repositories."""
2672
def _get_repo_format_to_test(self):
2673
from bzrlib.repofmt import pack_repo
2674
return pack_repo.RepositoryFormatKnitPack1()
2677
def is_compatible(source, target):
2678
"""Be compatible with known Pack formats.
2680
We don't test for the stores being of specific types because that
2681
could lead to confusing results, and there is no need to be
2684
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2686
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2687
isinstance(target._format, RepositoryFormatPack))
2688
except AttributeError:
2690
return are_packs and InterRepository._same_model(source, target)
2693
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2694
"""See InterRepository.fetch()."""
2695
from bzrlib.repofmt.pack_repo import Packer
2696
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2697
self.source, self.source._format, self.target, self.target._format)
2698
self.count_copied = 0
2699
if revision_id is None:
2701
# everything to do - use pack logic
2702
# to fetch from all packs to one without
2703
# inventory parsing etc, IFF nothing to be copied is in the target.
2705
revision_ids = self.source.all_revision_ids()
2706
revision_keys = [(revid,) for revid in revision_ids]
2707
index = self.target._pack_collection.revision_index.combined_index
2708
present_revision_ids = set(item[1][0] for item in
2709
index.iter_entries(revision_keys))
2710
revision_ids = set(revision_ids) - present_revision_ids
2711
# implementing the TODO will involve:
2712
# - detecting when all of a pack is selected
2713
# - avoiding as much as possible pre-selection, so the
2714
# more-core routines such as create_pack_from_packs can filter in
2715
# a just-in-time fashion. (though having a HEADS list on a
2716
# repository might make this a lot easier, because we could
2717
# sensibly detect 'new revisions' without doing a full index scan.
2718
elif _mod_revision.is_null(revision_id):
2723
revision_ids = self.search_missing_revision_ids(revision_id,
2724
find_ghosts=find_ghosts).get_keys()
2725
except errors.NoSuchRevision:
2726
raise errors.InstallFailed([revision_id])
2727
if len(revision_ids) == 0:
2729
packs = self.source._pack_collection.all_packs()
2730
pack = Packer(self.target._pack_collection, packs, '.fetch',
2731
revision_ids).pack()
2732
if pack is not None:
2733
self.target._pack_collection._save_pack_names()
2734
# Trigger an autopack. This may duplicate effort as we've just done
2735
# a pack creation, but for now it is simpler to think about as
2736
# 'upload data, then repack if needed'.
2737
self.target._pack_collection.autopack()
2738
return (pack.get_revision_count(), [])
2743
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2744
"""See InterRepository.missing_revision_ids().
2746
:param find_ghosts: Find ghosts throughout the ancestry of
2749
if not find_ghosts and revision_id is not None:
2750
return self._walk_to_common_revisions([revision_id])
2751
elif revision_id is not None:
2752
source_ids = self.source.get_ancestry(revision_id)
2753
if source_ids[0] is not None:
2754
raise AssertionError()
2757
source_ids = self.source.all_revision_ids()
2758
# source_ids is the worst possible case we may need to pull.
2759
# now we want to filter source_ids against what we actually
2760
# have in target, but don't try to check for existence where we know
2761
# we do not have a revision as that would be pointless.
2762
target_ids = set(self.target.all_revision_ids())
2763
result_set = set(source_ids).difference(target_ids)
2764
return self.source.revision_ids_to_search_result(result_set)
2767
class InterModel1and2(InterRepository):
2770
def _get_repo_format_to_test(self):
2774
def is_compatible(source, target):
2775
if not source.supports_rich_root() and target.supports_rich_root():
2781
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2782
"""See InterRepository.fetch()."""
2783
from bzrlib.fetch import Model1toKnit2Fetcher
2784
f = Model1toKnit2Fetcher(to_repository=self.target,
2785
from_repository=self.source,
2786
last_revision=revision_id,
2787
pb=pb, find_ghosts=find_ghosts)
2788
return f.count_copied, f.failed_revisions
2791
def copy_content(self, revision_id=None):
2792
"""Make a complete copy of the content in self into destination.
2794
This is a destructive operation! Do not use it on existing
2797
:param revision_id: Only copy the content needed to construct
2798
revision_id and its parents.
2801
self.target.set_make_working_trees(self.source.make_working_trees())
2802
except NotImplementedError:
2804
# but don't bother fetching if we have the needed data now.
2805
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2806
self.target.has_revision(revision_id)):
2808
self.target.fetch(self.source, revision_id=revision_id)
2811
class InterKnit1and2(InterKnitRepo):
2814
def _get_repo_format_to_test(self):
2818
def is_compatible(source, target):
2819
"""Be compatible with Knit1 source and Knit3 target"""
2820
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2822
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2823
RepositoryFormatKnit3)
2824
from bzrlib.repofmt.pack_repo import (
2825
RepositoryFormatKnitPack1,
2826
RepositoryFormatKnitPack3,
2827
RepositoryFormatPackDevelopment0,
2828
RepositoryFormatPackDevelopment0Subtree,
2831
RepositoryFormatKnit1,
2832
RepositoryFormatKnitPack1,
2833
RepositoryFormatPackDevelopment0,
2836
RepositoryFormatKnit3,
2837
RepositoryFormatKnitPack3,
2838
RepositoryFormatPackDevelopment0Subtree,
2840
return (isinstance(source._format, nosubtrees) and
2841
isinstance(target._format, subtrees))
2842
except AttributeError:
2846
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2847
"""See InterRepository.fetch()."""
2848
from bzrlib.fetch import Knit1to2Fetcher
2849
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2850
self.source, self.source._format, self.target,
2851
self.target._format)
2852
f = Knit1to2Fetcher(to_repository=self.target,
2853
from_repository=self.source,
2854
last_revision=revision_id,
2855
pb=pb, find_ghosts=find_ghosts)
2856
return f.count_copied, f.failed_revisions
2859
class InterDifferingSerializer(InterKnitRepo):
2862
def _get_repo_format_to_test(self):
2866
def is_compatible(source, target):
2867
"""Be compatible with Knit2 source and Knit3 target"""
2868
if source.supports_rich_root() != target.supports_rich_root():
2870
# Ideally, we'd support fetching if the source had no tree references
2871
# even if it supported them...
2872
if (getattr(source, '_format.supports_tree_reference', False) and
2873
not getattr(target, '_format.supports_tree_reference', False)):
2878
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2879
"""See InterRepository.fetch()."""
2880
revision_ids = self.target.search_missing_revision_ids(self.source,
2881
revision_id, find_ghosts=find_ghosts).get_keys()
2882
revision_ids = tsort.topo_sort(
2883
self.source.get_graph().get_parent_map(revision_ids))
2884
def revisions_iterator():
2885
for current_revision_id in revision_ids:
2886
revision = self.source.get_revision(current_revision_id)
2887
tree = self.source.revision_tree(current_revision_id)
2889
signature = self.source.get_signature_text(
2890
current_revision_id)
2891
except errors.NoSuchRevision:
2893
yield revision, tree, signature
2895
my_pb = ui.ui_factory.nested_progress_bar()
2900
install_revisions(self.target, revisions_iterator(),
2901
len(revision_ids), pb)
2903
if my_pb is not None:
2905
return len(revision_ids), 0
2908
class InterRemoteToOther(InterRepository):
2910
def __init__(self, source, target):
2911
InterRepository.__init__(self, source, target)
2912
self._real_inter = None
2915
def is_compatible(source, target):
2916
if not isinstance(source, remote.RemoteRepository):
2918
# Is source's model compatible with target's model?
2919
source._ensure_real()
2920
real_source = source._real_repository
2921
if isinstance(real_source, remote.RemoteRepository):
2922
raise NotImplementedError(
2923
"We don't support remote repos backed by remote repos yet.")
2924
return InterRepository._same_model(real_source, target)
2927
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2928
"""See InterRepository.fetch()."""
2929
from bzrlib.fetch import RemoteToOtherFetcher
2930
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2931
self.source, self.target, self.target._format)
2932
# TODO: jam 20070210 This should be an assert, not a translate
2933
revision_id = osutils.safe_revision_id(revision_id)
2934
f = RemoteToOtherFetcher(to_repository=self.target,
2935
from_repository=self.source,
2936
last_revision=revision_id,
2937
pb=pb, find_ghosts=find_ghosts)
2938
return f.count_copied, f.failed_revisions
2941
def _get_repo_format_to_test(self):
2945
class InterOtherToRemote(InterRepository):
2947
def __init__(self, source, target):
2948
InterRepository.__init__(self, source, target)
2949
self._real_inter = None
2952
def is_compatible(source, target):
2953
if isinstance(target, remote.RemoteRepository):
2957
def _ensure_real_inter(self):
2958
if self._real_inter is None:
2959
self.target._ensure_real()
2960
real_target = self.target._real_repository
2961
self._real_inter = InterRepository.get(self.source, real_target)
2963
def copy_content(self, revision_id=None):
2964
self._ensure_real_inter()
2965
self._real_inter.copy_content(revision_id=revision_id)
2967
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2968
self._ensure_real_inter()
2969
self._real_inter.fetch(revision_id=revision_id, pb=pb,
2970
find_ghosts=find_ghosts)
2973
def _get_repo_format_to_test(self):
2977
InterRepository.register_optimiser(InterDifferingSerializer)
2978
InterRepository.register_optimiser(InterSameDataRepository)
2979
InterRepository.register_optimiser(InterWeaveRepo)
2980
InterRepository.register_optimiser(InterKnitRepo)
2981
InterRepository.register_optimiser(InterModel1and2)
2982
InterRepository.register_optimiser(InterKnit1and2)
2983
InterRepository.register_optimiser(InterPackRepo)
2984
InterRepository.register_optimiser(InterRemoteToOther)
2985
InterRepository.register_optimiser(InterOtherToRemote)
2988
class CopyConverter(object):
2989
"""A repository conversion tool which just performs a copy of the content.
2991
This is slow but quite reliable.
2994
def __init__(self, target_format):
2995
"""Create a CopyConverter.
2997
:param target_format: The format the resulting repository should be.
2999
self.target_format = target_format
3001
def convert(self, repo, pb):
3002
"""Perform the conversion of to_convert, giving feedback via pb.
3004
:param to_convert: The disk object to convert.
3005
:param pb: a progress bar to use for progress information.
3010
# this is only useful with metadir layouts - separated repo content.
3011
# trigger an assertion if not such
3012
repo._format.get_format_string()
3013
self.repo_dir = repo.bzrdir
3014
self.step('Moving repository to repository.backup')
3015
self.repo_dir.transport.move('repository', 'repository.backup')
3016
backup_transport = self.repo_dir.transport.clone('repository.backup')
3017
repo._format.check_conversion_target(self.target_format)
3018
self.source_repo = repo._format.open(self.repo_dir,
3020
_override_transport=backup_transport)
3021
self.step('Creating new repository')
3022
converted = self.target_format.initialize(self.repo_dir,
3023
self.source_repo.is_shared())
3024
converted.lock_write()
3026
self.step('Copying content into repository.')
3027
self.source_repo.copy_content_into(converted)
3030
self.step('Deleting old repository content.')
3031
self.repo_dir.transport.delete_tree('repository.backup')
3032
self.pb.note('repository converted')
3034
def step(self, message):
3035
"""Update the pb by a step."""
3037
self.pb.update(message, self.count, self.total)
3049
def _unescaper(match, _map=_unescape_map):
3050
code = match.group(1)
3054
if not code.startswith('#'):
3056
return unichr(int(code[1:])).encode('utf8')
3062
def _unescape_xml(data):
3063
"""Unescape predefined XML entities in a string of data."""
3065
if _unescape_re is None:
3066
_unescape_re = re.compile('\&([^;]*);')
3067
return _unescape_re.sub(_unescaper, data)
3070
class _VersionedFileChecker(object):
3072
def __init__(self, repository):
3073
self.repository = repository
3074
self.text_index = self.repository._generate_text_key_index()
3076
def calculate_file_version_parents(self, revision_id, file_id):
3077
"""Calculate the correct parents for a file version according to
3080
parent_keys = self.text_index[(file_id, revision_id)]
3081
if parent_keys == [_mod_revision.NULL_REVISION]:
3083
# strip the file_id, for the weave api
3084
return tuple([revision_id for file_id, revision_id in parent_keys])
3086
def check_file_version_parents(self, weave, file_id):
3087
"""Check the parents stored in a versioned file are correct.
3089
It also detects file versions that are not referenced by their
3090
corresponding revision's inventory.
3092
:returns: A tuple of (wrong_parents, dangling_file_versions).
3093
wrong_parents is a dict mapping {revision_id: (stored_parents,
3094
correct_parents)} for each revision_id where the stored parents
3095
are not correct. dangling_file_versions is a set of (file_id,
3096
revision_id) tuples for versions that are present in this versioned
3097
file, but not used by the corresponding inventory.
3100
unused_versions = set()
3101
versions = weave.versions()
3102
parent_map = weave.get_parent_map(versions)
3103
for num, revision_id in enumerate(versions):
3105
correct_parents = self.calculate_file_version_parents(
3106
revision_id, file_id)
3108
# The version is not part of the used keys.
3109
unused_versions.add(revision_id)
3112
knit_parents = tuple(parent_map[revision_id])
3113
except errors.RevisionNotPresent:
3115
if correct_parents != knit_parents:
3116
wrong_parents[revision_id] = (knit_parents, correct_parents)
3117
return wrong_parents, unused_versions
3120
def _old_get_graph(repository, revision_id):
3121
"""DO NOT USE. That is all. I'm serious."""
3122
graph = repository.get_graph()
3123
revision_graph = dict(((key, value) for key, value in
3124
graph.iter_ancestry([revision_id]) if value is not None))
3125
return _strip_NULL_ghosts(revision_graph)
3128
def _strip_NULL_ghosts(revision_graph):
3129
"""Also don't use this. more compatibility code for unmigrated clients."""
3130
# Filter ghosts, and null:
3131
if _mod_revision.NULL_REVISION in revision_graph:
3132
del revision_graph[_mod_revision.NULL_REVISION]
3133
for key, parents in revision_graph.items():
3134
revision_graph[key] = tuple(parent for parent in parents if parent
3136
return revision_graph
added
removed
bzrlib/repository.py
