1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
39
revision as _mod_revision,
45
from bzrlib.bundle import serializer
46
from bzrlib.revisiontree import RevisionTree
47
from bzrlib.store.versioned import VersionedFileStore
48
from bzrlib.store.text import TextStore
49
from bzrlib.testament import Testament
50
from bzrlib.util import bencode
53
from bzrlib.decorators import needs_read_lock, needs_write_lock
54
from bzrlib.inter import InterObject
55
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
56
from bzrlib.symbol_versioning import (
59
from bzrlib.trace import mutter, mutter_callsite, note, warning
62
# Old formats display a warning, but only once
63
_deprecation_warning_done = False
66
class CommitBuilder(object):
67
"""Provides an interface to build up a commit.
69
This allows describing a tree to be committed without needing to
70
know the internals of the format of the repository.
73
# all clients should supply tree roots.
74
record_root_entry = True
75
# the default CommitBuilder does not manage trees whose root is versioned.
76
_versioned_root = False
78
def __init__(self, repository, parents, config, timestamp=None,
79
timezone=None, committer=None, revprops=None,
81
"""Initiate a CommitBuilder.
83
:param repository: Repository to commit to.
84
:param parents: Revision ids of the parents of the new revision.
85
:param config: Configuration to use.
86
:param timestamp: Optional timestamp recorded for commit.
87
:param timezone: Optional timezone for timestamp.
88
:param committer: Optional committer to set for commit.
89
:param revprops: Optional dictionary of revision properties.
90
:param revision_id: Optional revision id.
95
self._committer = self._config.username()
97
self._committer = committer
99
self.new_inventory = Inventory(None)
100
self._new_revision_id = revision_id
101
self.parents = parents
102
self.repository = repository
105
if revprops is not None:
106
self._revprops.update(revprops)
108
if timestamp is None:
109
timestamp = time.time()
110
# Restrict resolution to 1ms
111
self._timestamp = round(timestamp, 3)
114
self._timezone = osutils.local_time_offset()
116
self._timezone = int(timezone)
118
self._generate_revision_if_needed()
119
self.__heads = graph.HeadsCache(repository.get_graph()).heads
121
def commit(self, message):
122
"""Make the actual commit.
124
:return: The revision id of the recorded revision.
126
rev = _mod_revision.Revision(
127
timestamp=self._timestamp,
128
timezone=self._timezone,
129
committer=self._committer,
131
inventory_sha1=self.inv_sha1,
132
revision_id=self._new_revision_id,
133
properties=self._revprops)
134
rev.parent_ids = self.parents
135
self.repository.add_revision(self._new_revision_id, rev,
136
self.new_inventory, self._config)
137
self.repository.commit_write_group()
138
return self._new_revision_id
141
"""Abort the commit that is being built.
143
self.repository.abort_write_group()
145
def revision_tree(self):
146
"""Return the tree that was just committed.
148
After calling commit() this can be called to get a RevisionTree
149
representing the newly committed tree. This is preferred to
150
calling Repository.revision_tree() because that may require
151
deserializing the inventory, while we already have a copy in
154
return RevisionTree(self.repository, self.new_inventory,
155
self._new_revision_id)
157
def finish_inventory(self):
158
"""Tell the builder that the inventory is finished."""
159
if self.new_inventory.root is None:
160
raise AssertionError('Root entry should be supplied to'
161
' record_entry_contents, as of bzr 0.10.',
162
DeprecationWarning, stacklevel=2)
163
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
164
self.new_inventory.revision_id = self._new_revision_id
165
self.inv_sha1 = self.repository.add_inventory(
166
self._new_revision_id,
171
def _gen_revision_id(self):
172
"""Return new revision-id."""
173
return generate_ids.gen_revision_id(self._config.username(),
176
def _generate_revision_if_needed(self):
177
"""Create a revision id if None was supplied.
179
If the repository can not support user-specified revision ids
180
they should override this function and raise CannotSetRevisionId
181
if _new_revision_id is not None.
183
:raises: CannotSetRevisionId
185
if self._new_revision_id is None:
186
self._new_revision_id = self._gen_revision_id()
187
self.random_revid = True
189
self.random_revid = False
191
def _heads(self, file_id, revision_ids):
192
"""Calculate the graph heads for revision_ids in the graph of file_id.
194
This can use either a per-file graph or a global revision graph as we
195
have an identity relationship between the two graphs.
197
return self.__heads(revision_ids)
199
def _check_root(self, ie, parent_invs, tree):
200
"""Helper for record_entry_contents.
202
:param ie: An entry being added.
203
:param parent_invs: The inventories of the parent revisions of the
205
:param tree: The tree that is being committed.
207
# In this revision format, root entries have no knit or weave When
208
# serializing out to disk and back in root.revision is always
210
ie.revision = self._new_revision_id
212
def _get_delta(self, ie, basis_inv, path):
213
"""Get a delta against the basis inventory for ie."""
214
if ie.file_id not in basis_inv:
216
return (None, path, ie.file_id, ie)
217
elif ie != basis_inv[ie.file_id]:
219
# TODO: avoid tis id2path call.
220
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
225
def record_entry_contents(self, ie, parent_invs, path, tree,
227
"""Record the content of ie from tree into the commit if needed.
229
Side effect: sets ie.revision when unchanged
231
:param ie: An inventory entry present in the commit.
232
:param parent_invs: The inventories of the parent revisions of the
234
:param path: The path the entry is at in the tree.
235
:param tree: The tree which contains this entry and should be used to
237
:param content_summary: Summary data from the tree about the paths
238
content - stat, length, exec, sha/link target. This is only
239
accessed when the entry has a revision of None - that is when it is
240
a candidate to commit.
241
:return: A tuple (change_delta, version_recorded). change_delta is
242
an inventory_delta change for this entry against the basis tree of
243
the commit, or None if no change occured against the basis tree.
244
version_recorded is True if a new version of the entry has been
245
recorded. For instance, committing a merge where a file was only
246
changed on the other side will return (delta, False).
248
if self.new_inventory.root is None:
249
if ie.parent_id is not None:
250
raise errors.RootMissing()
251
self._check_root(ie, parent_invs, tree)
252
if ie.revision is None:
253
kind = content_summary[0]
255
# ie is carried over from a prior commit
257
# XXX: repository specific check for nested tree support goes here - if
258
# the repo doesn't want nested trees we skip it ?
259
if (kind == 'tree-reference' and
260
not self.repository._format.supports_tree_reference):
261
# mismatch between commit builder logic and repository:
262
# this needs the entry creation pushed down into the builder.
263
raise NotImplementedError('Missing repository subtree support.')
264
self.new_inventory.add(ie)
266
# TODO: slow, take it out of the inner loop.
268
basis_inv = parent_invs[0]
270
basis_inv = Inventory(root_id=None)
272
# ie.revision is always None if the InventoryEntry is considered
273
# for committing. We may record the previous parents revision if the
274
# content is actually unchanged against a sole head.
275
if ie.revision is not None:
276
if not self._versioned_root and path == '':
277
# repositories that do not version the root set the root's
278
# revision to the new commit even when no change occurs, and
279
# this masks when a change may have occurred against the basis,
280
# so calculate if one happened.
281
if ie.file_id in basis_inv:
282
delta = (basis_inv.id2path(ie.file_id), path,
286
delta = (None, path, ie.file_id, ie)
289
# we don't need to commit this, because the caller already
290
# determined that an existing revision of this file is
292
return None, (ie.revision == self._new_revision_id)
293
# XXX: Friction: parent_candidates should return a list not a dict
294
# so that we don't have to walk the inventories again.
295
parent_candiate_entries = ie.parent_candidates(parent_invs)
296
head_set = self._heads(ie.file_id, parent_candiate_entries.keys())
298
for inv in parent_invs:
299
if ie.file_id in inv:
300
old_rev = inv[ie.file_id].revision
301
if old_rev in head_set:
302
heads.append(inv[ie.file_id].revision)
303
head_set.remove(inv[ie.file_id].revision)
306
# now we check to see if we need to write a new record to the
308
# We write a new entry unless there is one head to the ancestors, and
309
# the kind-derived content is unchanged.
311
# Cheapest check first: no ancestors, or more the one head in the
312
# ancestors, we write a new node.
316
# There is a single head, look it up for comparison
317
parent_entry = parent_candiate_entries[heads[0]]
318
# if the non-content specific data has changed, we'll be writing a
320
if (parent_entry.parent_id != ie.parent_id or
321
parent_entry.name != ie.name):
323
# now we need to do content specific checks:
325
# if the kind changed the content obviously has
326
if kind != parent_entry.kind:
329
if content_summary[2] is None:
330
raise ValueError("Files must not have executable = None")
332
if (# if the file length changed we have to store:
333
parent_entry.text_size != content_summary[1] or
334
# if the exec bit has changed we have to store:
335
parent_entry.executable != content_summary[2]):
337
elif parent_entry.text_sha1 == content_summary[3]:
338
# all meta and content is unchanged (using a hash cache
339
# hit to check the sha)
340
ie.revision = parent_entry.revision
341
ie.text_size = parent_entry.text_size
342
ie.text_sha1 = parent_entry.text_sha1
343
ie.executable = parent_entry.executable
344
return self._get_delta(ie, basis_inv, path), False
346
# Either there is only a hash change(no hash cache entry,
347
# or same size content change), or there is no change on
349
# Provide the parent's hash to the store layer, so that the
350
# content is unchanged we will not store a new node.
351
nostore_sha = parent_entry.text_sha1
353
# We want to record a new node regardless of the presence or
354
# absence of a content change in the file.
356
ie.executable = content_summary[2]
357
lines = tree.get_file(ie.file_id, path).readlines()
359
ie.text_sha1, ie.text_size = self._add_text_to_weave(
360
ie.file_id, lines, heads, nostore_sha)
361
except errors.ExistingContent:
362
# Turns out that the file content was unchanged, and we were
363
# only going to store a new node if it was changed. Carry over
365
ie.revision = parent_entry.revision
366
ie.text_size = parent_entry.text_size
367
ie.text_sha1 = parent_entry.text_sha1
368
ie.executable = parent_entry.executable
369
return self._get_delta(ie, basis_inv, path), False
370
elif kind == 'directory':
372
# all data is meta here, nothing specific to directory, so
374
ie.revision = parent_entry.revision
375
return self._get_delta(ie, basis_inv, path), False
377
self._add_text_to_weave(ie.file_id, lines, heads, None)
378
elif kind == 'symlink':
379
current_link_target = content_summary[3]
381
# symlink target is not generic metadata, check if it has
383
if current_link_target != parent_entry.symlink_target:
386
# unchanged, carry over.
387
ie.revision = parent_entry.revision
388
ie.symlink_target = parent_entry.symlink_target
389
return self._get_delta(ie, basis_inv, path), False
390
ie.symlink_target = current_link_target
392
self._add_text_to_weave(ie.file_id, lines, heads, None)
393
elif kind == 'tree-reference':
395
if content_summary[3] != parent_entry.reference_revision:
398
# unchanged, carry over.
399
ie.reference_revision = parent_entry.reference_revision
400
ie.revision = parent_entry.revision
401
return self._get_delta(ie, basis_inv, path), False
402
ie.reference_revision = content_summary[3]
404
self._add_text_to_weave(ie.file_id, lines, heads, None)
406
raise NotImplementedError('unknown kind')
407
ie.revision = self._new_revision_id
408
return self._get_delta(ie, basis_inv, path), True
410
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
411
versionedfile = self.repository.weave_store.get_weave_or_empty(
412
file_id, self.repository.get_transaction())
413
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
414
# than add_lines, and allows committing when a parent is ghosted for
416
# Note: as we read the content directly from the tree, we know its not
417
# been turned into unicode or badly split - but a broken tree
418
# implementation could give us bad output from readlines() so this is
419
# not a guarantee of safety. What would be better is always checking
420
# the content during test suite execution. RBC 20070912
421
return versionedfile.add_lines_with_ghosts(
422
self._new_revision_id, parents, new_lines,
423
nostore_sha=nostore_sha, random_id=self.random_revid,
424
check_content=False)[0:2]
427
class RootCommitBuilder(CommitBuilder):
428
"""This commitbuilder actually records the root id"""
430
# the root entry gets versioned properly by this builder.
431
_versioned_root = True
433
def _check_root(self, ie, parent_invs, tree):
434
"""Helper for record_entry_contents.
436
:param ie: An entry being added.
437
:param parent_invs: The inventories of the parent revisions of the
439
:param tree: The tree that is being committed.
443
######################################################################
446
class Repository(object):
447
"""Repository holding history for one or more branches.
449
The repository holds and retrieves historical information including
450
revisions and file history. It's normally accessed only by the Branch,
451
which views a particular line of development through that history.
453
The Repository builds on top of Stores and a Transport, which respectively
454
describe the disk data format and the way of accessing the (possibly
457
:ivar _transport: Transport for file access to repository, typically
458
pointing to .bzr/repository.
461
# What class to use for a CommitBuilder. Often its simpler to change this
462
# in a Repository class subclass rather than to override
463
# get_commit_builder.
464
_commit_builder_class = CommitBuilder
465
# The search regex used by xml based repositories to determine what things
466
# where changed in a single commit.
467
_file_ids_altered_regex = lazy_regex.lazy_compile(
468
r'file_id="(?P<file_id>[^"]+)"'
469
r'.* revision="(?P<revision_id>[^"]+)"'
472
def abort_write_group(self):
473
"""Commit the contents accrued within the current write group.
475
:seealso: start_write_group.
477
if self._write_group is not self.get_transaction():
478
# has an unlock or relock occured ?
479
raise errors.BzrError('mismatched lock context and write group.')
480
self._abort_write_group()
481
self._write_group = None
483
def _abort_write_group(self):
484
"""Template method for per-repository write group cleanup.
486
This is called during abort before the write group is considered to be
487
finished and should cleanup any internal state accrued during the write
488
group. There is no requirement that data handed to the repository be
489
*not* made available - this is not a rollback - but neither should any
490
attempt be made to ensure that data added is fully commited. Abort is
491
invoked when an error has occured so futher disk or network operations
492
may not be possible or may error and if possible should not be
496
def add_inventory(self, revision_id, inv, parents):
497
"""Add the inventory inv to the repository as revision_id.
499
:param parents: The revision ids of the parents that revision_id
500
is known to have and are in the repository already.
502
:returns: The validator(which is a sha1 digest, though what is sha'd is
503
repository format specific) of the serialized inventory.
505
if not self.is_in_write_group():
506
raise AssertionError("%r not in write group" % (self,))
507
_mod_revision.check_not_reserved_id(revision_id)
508
if not (inv.revision_id is None or inv.revision_id == revision_id):
509
raise AssertionError(
510
"Mismatch between inventory revision"
511
" id and insertion revid (%r, %r)"
512
% (inv.revision_id, revision_id))
514
raise AssertionError()
515
inv_lines = self._serialise_inventory_to_lines(inv)
516
inv_vf = self.get_inventory_weave()
517
return self._inventory_add_lines(inv_vf, revision_id, parents,
518
inv_lines, check_content=False)
520
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
522
"""Store lines in inv_vf and return the sha1 of the inventory."""
524
for parent in parents:
526
final_parents.append(parent)
527
return inv_vf.add_lines(revision_id, final_parents, lines,
528
check_content=check_content)[0]
530
def add_revision(self, revision_id, rev, inv=None, config=None):
531
"""Add rev to the revision store as revision_id.
533
:param revision_id: the revision id to use.
534
:param rev: The revision object.
535
:param inv: The inventory for the revision. if None, it will be looked
536
up in the inventory storer
537
:param config: If None no digital signature will be created.
538
If supplied its signature_needed method will be used
539
to determine if a signature should be made.
541
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
543
_mod_revision.check_not_reserved_id(revision_id)
544
if config is not None and config.signature_needed():
546
inv = self.get_inventory(revision_id)
547
plaintext = Testament(rev, inv).as_short_text()
548
self.store_revision_signature(
549
gpg.GPGStrategy(config), plaintext, revision_id)
550
inventory_vf = self.get_inventory_weave()
551
if not revision_id in inventory_vf:
553
raise errors.WeaveRevisionNotPresent(revision_id,
556
# yes, this is not suitable for adding with ghosts.
557
rev.inventory_sha1 = self.add_inventory(revision_id, inv,
560
rev.inventory_sha1 = inventory_vf.get_sha1s([revision_id])[0]
561
self._revision_store.add_revision(rev, self.get_transaction())
563
def _add_revision_text(self, revision_id, text):
564
revision = self._revision_store._serializer.read_revision_from_string(
566
self._revision_store._add_revision(revision, StringIO(text),
567
self.get_transaction())
569
def all_revision_ids(self):
570
"""Returns a list of all the revision ids in the repository.
572
This is deprecated because code should generally work on the graph
573
reachable from a particular revision, and ignore any other revisions
574
that might be present. There is no direct replacement method.
576
if 'evil' in debug.debug_flags:
577
mutter_callsite(2, "all_revision_ids is linear with history.")
578
return self._all_revision_ids()
580
def _all_revision_ids(self):
581
"""Returns a list of all the revision ids in the repository.
583
These are in as much topological order as the underlying store can
586
raise NotImplementedError(self._all_revision_ids)
588
def break_lock(self):
589
"""Break a lock if one is present from another instance.
591
Uses the ui factory to ask for confirmation if the lock may be from
594
self.control_files.break_lock()
597
def _eliminate_revisions_not_present(self, revision_ids):
598
"""Check every revision id in revision_ids to see if we have it.
600
Returns a set of the present revisions.
603
graph = self.get_graph()
604
parent_map = graph.get_parent_map(revision_ids)
605
# The old API returned a list, should this actually be a set?
606
return parent_map.keys()
609
def create(a_bzrdir):
610
"""Construct the current default format repository in a_bzrdir."""
611
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
613
def __init__(self, _format, a_bzrdir, control_files,
614
_revision_store, control_store, text_store):
615
"""instantiate a Repository.
617
:param _format: The format of the repository on disk.
618
:param a_bzrdir: The BzrDir of the repository.
620
In the future we will have a single api for all stores for
621
getting file texts, inventories and revisions, then
622
this construct will accept instances of those things.
624
super(Repository, self).__init__()
625
self._format = _format
626
# the following are part of the public API for Repository:
627
self.bzrdir = a_bzrdir
628
self.control_files = control_files
629
self._transport = control_files._transport
630
self.base = self._transport.base
631
self._revision_store = _revision_store
632
# backwards compatibility
633
self.weave_store = text_store
635
self._reconcile_does_inventory_gc = True
636
self._reconcile_fixes_text_parents = False
637
self._reconcile_backsup_inventory = True
638
# not right yet - should be more semantically clear ?
640
self.control_store = control_store
641
self.control_weaves = control_store
642
# TODO: make sure to construct the right store classes, etc, depending
643
# on whether escaping is required.
644
self._warn_if_deprecated()
645
self._write_group = None
648
return '%s(%r)' % (self.__class__.__name__,
651
def has_same_location(self, other):
652
"""Returns a boolean indicating if this repository is at the same
653
location as another repository.
655
This might return False even when two repository objects are accessing
656
the same physical repository via different URLs.
658
if self.__class__ is not other.__class__:
660
return (self._transport.base == other._transport.base)
662
def is_in_write_group(self):
663
"""Return True if there is an open write group.
665
:seealso: start_write_group.
667
return self._write_group is not None
670
return self.control_files.is_locked()
672
def is_write_locked(self):
673
"""Return True if this object is write locked."""
674
return self.is_locked() and self.control_files._lock_mode == 'w'
676
def lock_write(self, token=None):
677
"""Lock this repository for writing.
679
This causes caching within the repository obejct to start accumlating
680
data during reads, and allows a 'write_group' to be obtained. Write
681
groups must be used for actual data insertion.
683
:param token: if this is already locked, then lock_write will fail
684
unless the token matches the existing lock.
685
:returns: a token if this instance supports tokens, otherwise None.
686
:raises TokenLockingNotSupported: when a token is given but this
687
instance doesn't support using token locks.
688
:raises MismatchedToken: if the specified token doesn't match the token
689
of the existing lock.
690
:seealso: start_write_group.
692
A token should be passed in if you know that you have locked the object
693
some other way, and need to synchronise this object's state with that
696
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
698
result = self.control_files.lock_write(token=token)
703
self.control_files.lock_read()
706
def get_physical_lock_status(self):
707
return self.control_files.get_physical_lock_status()
709
def leave_lock_in_place(self):
710
"""Tell this repository not to release the physical lock when this
713
If lock_write doesn't return a token, then this method is not supported.
715
self.control_files.leave_in_place()
717
def dont_leave_lock_in_place(self):
718
"""Tell this repository to release the physical lock when this
719
object is unlocked, even if it didn't originally acquire it.
721
If lock_write doesn't return a token, then this method is not supported.
723
self.control_files.dont_leave_in_place()
726
def gather_stats(self, revid=None, committers=None):
727
"""Gather statistics from a revision id.
729
:param revid: The revision id to gather statistics from, if None, then
730
no revision specific statistics are gathered.
731
:param committers: Optional parameter controlling whether to grab
732
a count of committers from the revision specific statistics.
733
:return: A dictionary of statistics. Currently this contains:
734
committers: The number of committers if requested.
735
firstrev: A tuple with timestamp, timezone for the penultimate left
736
most ancestor of revid, if revid is not the NULL_REVISION.
737
latestrev: A tuple with timestamp, timezone for revid, if revid is
738
not the NULL_REVISION.
739
revisions: The total revision count in the repository.
740
size: An estimate disk size of the repository in bytes.
743
if revid and committers:
744
result['committers'] = 0
745
if revid and revid != _mod_revision.NULL_REVISION:
747
all_committers = set()
748
revisions = self.get_ancestry(revid)
749
# pop the leading None
751
first_revision = None
753
# ignore the revisions in the middle - just grab first and last
754
revisions = revisions[0], revisions[-1]
755
for revision in self.get_revisions(revisions):
756
if not first_revision:
757
first_revision = revision
759
all_committers.add(revision.committer)
760
last_revision = revision
762
result['committers'] = len(all_committers)
763
result['firstrev'] = (first_revision.timestamp,
764
first_revision.timezone)
765
result['latestrev'] = (last_revision.timestamp,
766
last_revision.timezone)
768
# now gather global repository information
769
if self.bzrdir.root_transport.listable():
770
c, t = self._revision_store.total_size(self.get_transaction())
771
result['revisions'] = c
775
def find_branches(self, using=False):
776
"""Find branches underneath this repository.
778
This will include branches inside other branches.
780
:param using: If True, list only branches using this repository.
782
if using and not self.is_shared():
784
return [self.bzrdir.open_branch()]
785
except errors.NotBranchError:
787
class Evaluator(object):
790
self.first_call = True
792
def __call__(self, bzrdir):
793
# On the first call, the parameter is always the bzrdir
794
# containing the current repo.
795
if not self.first_call:
797
repository = bzrdir.open_repository()
798
except errors.NoRepositoryPresent:
801
return False, (None, repository)
802
self.first_call = False
804
value = (bzrdir.open_branch(), None)
805
except errors.NotBranchError:
810
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
811
self.bzrdir.root_transport, evaluate=Evaluator()):
812
if branch is not None:
813
branches.append(branch)
814
if not using and repository is not None:
815
branches.extend(repository.find_branches())
818
def get_data_stream(self, revision_ids):
819
raise NotImplementedError(self.get_data_stream)
821
def get_data_stream_for_search(self, search_result):
822
"""Get a data stream that can be inserted to a repository.
824
:param search_result: A bzrlib.graph.SearchResult selecting the
826
:return: A data stream that can be inserted into a repository using
829
raise NotImplementedError(self.get_data_stream_for_search)
831
def insert_data_stream(self, stream):
832
"""XXX What does this really do?
834
Is it a substitute for fetch?
835
Should it manage its own write group ?
837
for item_key, bytes in stream:
838
if item_key[0] == 'file':
839
(file_id,) = item_key[1:]
840
knit = self.weave_store.get_weave_or_empty(
841
file_id, self.get_transaction())
842
elif item_key == ('inventory',):
843
knit = self.get_inventory_weave()
844
elif item_key == ('revisions',):
845
knit = self._revision_store.get_revision_file(
846
self.get_transaction())
847
elif item_key == ('signatures',):
848
knit = self._revision_store.get_signature_file(
849
self.get_transaction())
851
raise errors.RepositoryDataStreamError(
852
"Unrecognised data stream key '%s'" % (item_key,))
853
decoded_list = bencode.bdecode(bytes)
854
format = decoded_list.pop(0)
857
for version, options, parents, some_bytes in decoded_list:
858
data_list.append((version, options, len(some_bytes), parents))
859
knit_bytes += some_bytes
860
buffer = StringIO(knit_bytes)
861
def reader_func(count):
865
return buffer.read(count)
866
knit.insert_data_stream(
867
(format, data_list, reader_func))
870
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
871
"""Return the revision ids that other has that this does not.
873
These are returned in topological order.
875
revision_id: only return revision ids included by revision_id.
877
return InterRepository.get(other, self).search_missing_revision_ids(
878
revision_id, find_ghosts)
880
@deprecated_method(symbol_versioning.one_two)
882
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
883
"""Return the revision ids that other has that this does not.
885
These are returned in topological order.
887
revision_id: only return revision ids included by revision_id.
889
keys = self.search_missing_revision_ids(
890
other, revision_id, find_ghosts).get_keys()
893
parents = other.get_graph().get_parent_map(keys)
896
return tsort.topo_sort(parents)
900
"""Open the repository rooted at base.
902
For instance, if the repository is at URL/.bzr/repository,
903
Repository.open(URL) -> a Repository instance.
905
control = bzrdir.BzrDir.open(base)
906
return control.open_repository()
908
def copy_content_into(self, destination, revision_id=None):
909
"""Make a complete copy of the content in self into destination.
911
This is a destructive operation! Do not use it on existing
914
return InterRepository.get(self, destination).copy_content(revision_id)
916
def commit_write_group(self):
917
"""Commit the contents accrued within the current write group.
919
:seealso: start_write_group.
921
if self._write_group is not self.get_transaction():
922
# has an unlock or relock occured ?
923
raise errors.BzrError('mismatched lock context %r and '
925
(self.get_transaction(), self._write_group))
926
self._commit_write_group()
927
self._write_group = None
929
def _commit_write_group(self):
930
"""Template method for per-repository write group cleanup.
932
This is called before the write group is considered to be
933
finished and should ensure that all data handed to the repository
934
for writing during the write group is safely committed (to the
935
extent possible considering file system caching etc).
938
def fetch(self, source, revision_id=None, pb=None, find_ghosts=False):
939
"""Fetch the content required to construct revision_id from source.
941
If revision_id is None all content is copied.
942
:param find_ghosts: Find and copy revisions in the source that are
943
ghosts in the target (and not reachable directly by walking out to
944
the first-present revision in target from revision_id).
946
# fast path same-url fetch operations
947
if self.has_same_location(source):
948
# check that last_revision is in 'from' and then return a
950
if (revision_id is not None and
951
not _mod_revision.is_null(revision_id)):
952
self.get_revision(revision_id)
954
inter = InterRepository.get(source, self)
956
return inter.fetch(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts)
957
except NotImplementedError:
958
raise errors.IncompatibleRepositories(source, self)
960
def create_bundle(self, target, base, fileobj, format=None):
961
return serializer.write_bundle(self, target, base, fileobj, format)
963
def get_commit_builder(self, branch, parents, config, timestamp=None,
964
timezone=None, committer=None, revprops=None,
966
"""Obtain a CommitBuilder for this repository.
968
:param branch: Branch to commit to.
969
:param parents: Revision ids of the parents of the new revision.
970
:param config: Configuration to use.
971
:param timestamp: Optional timestamp recorded for commit.
972
:param timezone: Optional timezone for timestamp.
973
:param committer: Optional committer to set for commit.
974
:param revprops: Optional dictionary of revision properties.
975
:param revision_id: Optional revision id.
977
result = self._commit_builder_class(self, parents, config,
978
timestamp, timezone, committer, revprops, revision_id)
979
self.start_write_group()
983
if (self.control_files._lock_count == 1 and
984
self.control_files._lock_mode == 'w'):
985
if self._write_group is not None:
986
self.abort_write_group()
987
self.control_files.unlock()
988
raise errors.BzrError(
989
'Must end write groups before releasing write locks.')
990
self.control_files.unlock()
993
def clone(self, a_bzrdir, revision_id=None):
994
"""Clone this repository into a_bzrdir using the current format.
996
Currently no check is made that the format of this repository and
997
the bzrdir format are compatible. FIXME RBC 20060201.
999
:return: The newly created destination repository.
1001
# TODO: deprecate after 0.16; cloning this with all its settings is
1002
# probably not very useful -- mbp 20070423
1003
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
1004
self.copy_content_into(dest_repo, revision_id)
1007
def start_write_group(self):
1008
"""Start a write group in the repository.
1010
Write groups are used by repositories which do not have a 1:1 mapping
1011
between file ids and backend store to manage the insertion of data from
1012
both fetch and commit operations.
1014
A write lock is required around the start_write_group/commit_write_group
1015
for the support of lock-requiring repository formats.
1017
One can only insert data into a repository inside a write group.
1021
if not self.is_write_locked():
1022
raise errors.NotWriteLocked(self)
1023
if self._write_group:
1024
raise errors.BzrError('already in a write group')
1025
self._start_write_group()
1026
# so we can detect unlock/relock - the write group is now entered.
1027
self._write_group = self.get_transaction()
1029
def _start_write_group(self):
1030
"""Template method for per-repository write group startup.
1032
This is called before the write group is considered to be
1037
def sprout(self, to_bzrdir, revision_id=None):
1038
"""Create a descendent repository for new development.
1040
Unlike clone, this does not copy the settings of the repository.
1042
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
1043
dest_repo.fetch(self, revision_id=revision_id)
1046
def _create_sprouting_repo(self, a_bzrdir, shared):
1047
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
1048
# use target default format.
1049
dest_repo = a_bzrdir.create_repository()
1051
# Most control formats need the repository to be specifically
1052
# created, but on some old all-in-one formats it's not needed
1054
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
1055
except errors.UninitializableFormat:
1056
dest_repo = a_bzrdir.open_repository()
1060
def has_revision(self, revision_id):
1061
"""True if this repository has a copy of the revision."""
1062
return revision_id in self.has_revisions((revision_id,))
1064
def has_revisions(self, revision_ids):
1065
"""Probe to find out the presence of multiple revisions.
1067
:param revision_ids: An iterable of revision_ids.
1068
:return: A set of the revision_ids that were present.
1070
raise NotImplementedError(self.has_revisions)
1072
return self._revision_store.has_revision_id(revision_id,
1073
self.get_transaction())
1076
def get_revision(self, revision_id):
1077
"""Return the Revision object for a named revision."""
1078
return self.get_revisions([revision_id])[0]
1081
def get_revision_reconcile(self, revision_id):
1082
"""'reconcile' helper routine that allows access to a revision always.
1084
This variant of get_revision does not cross check the weave graph
1085
against the revision one as get_revision does: but it should only
1086
be used by reconcile, or reconcile-alike commands that are correcting
1087
or testing the revision graph.
1089
return self._get_revisions([revision_id])[0]
1092
def get_revisions(self, revision_ids):
1093
"""Get many revisions at once."""
1094
return self._get_revisions(revision_ids)
1097
def _get_revisions(self, revision_ids):
1098
"""Core work logic to get many revisions without sanity checks."""
1099
for rev_id in revision_ids:
1100
if not rev_id or not isinstance(rev_id, basestring):
1101
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1102
revs = self._revision_store.get_revisions(revision_ids,
1103
self.get_transaction())
1107
def get_revision_xml(self, revision_id):
1108
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1109
# would have already do it.
1110
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1111
rev = self.get_revision(revision_id)
1112
rev_tmp = StringIO()
1113
# the current serializer..
1114
self._revision_store._serializer.write_revision(rev, rev_tmp)
1116
return rev_tmp.getvalue()
1118
def get_deltas_for_revisions(self, revisions):
1119
"""Produce a generator of revision deltas.
1121
Note that the input is a sequence of REVISIONS, not revision_ids.
1122
Trees will be held in memory until the generator exits.
1123
Each delta is relative to the revision's lefthand predecessor.
1125
required_trees = set()
1126
for revision in revisions:
1127
required_trees.add(revision.revision_id)
1128
required_trees.update(revision.parent_ids[:1])
1129
trees = dict((t.get_revision_id(), t) for
1130
t in self.revision_trees(required_trees))
1131
for revision in revisions:
1132
if not revision.parent_ids:
1133
old_tree = self.revision_tree(None)
1135
old_tree = trees[revision.parent_ids[0]]
1136
yield trees[revision.revision_id].changes_from(old_tree)
1139
def get_revision_delta(self, revision_id):
1140
"""Return the delta for one revision.
1142
The delta is relative to the left-hand predecessor of the
1145
r = self.get_revision(revision_id)
1146
return list(self.get_deltas_for_revisions([r]))[0]
1149
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1150
signature = gpg_strategy.sign(plaintext)
1151
self.add_signature_text(revision_id, signature)
1154
def add_signature_text(self, revision_id, signature):
1155
self._revision_store.add_revision_signature_text(revision_id,
1157
self.get_transaction())
1159
def find_text_key_references(self):
1160
"""Find the text key references within the repository.
1162
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1163
revision_ids. Each altered file-ids has the exact revision_ids that
1164
altered it listed explicitly.
1165
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1166
to whether they were referred to by the inventory of the
1167
revision_id that they contain. The inventory texts from all present
1168
revision ids are assessed to generate this report.
1170
revision_ids = self.all_revision_ids()
1171
w = self.get_inventory_weave()
1172
pb = ui.ui_factory.nested_progress_bar()
1174
return self._find_text_key_references_from_xml_inventory_lines(
1175
w.iter_lines_added_or_present_in_versions(revision_ids, pb=pb))
1179
def _find_text_key_references_from_xml_inventory_lines(self,
1181
"""Core routine for extracting references to texts from inventories.
1183
This performs the translation of xml lines to revision ids.
1185
:param line_iterator: An iterator of lines, origin_version_id
1186
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1187
to whether they were referred to by the inventory of the
1188
revision_id that they contain. Note that if that revision_id was
1189
not part of the line_iterator's output then False will be given -
1190
even though it may actually refer to that key.
1192
if not self._serializer.support_altered_by_hack:
1193
raise AssertionError(
1194
"_find_text_key_references_from_xml_inventory_lines only "
1195
"supported for branches which store inventory as unnested xml"
1196
", not on %r" % self)
1199
# this code needs to read every new line in every inventory for the
1200
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1201
# not present in one of those inventories is unnecessary but not
1202
# harmful because we are filtering by the revision id marker in the
1203
# inventory lines : we only select file ids altered in one of those
1204
# revisions. We don't need to see all lines in the inventory because
1205
# only those added in an inventory in rev X can contain a revision=X
1207
unescape_revid_cache = {}
1208
unescape_fileid_cache = {}
1210
# jam 20061218 In a big fetch, this handles hundreds of thousands
1211
# of lines, so it has had a lot of inlining and optimizing done.
1212
# Sorry that it is a little bit messy.
1213
# Move several functions to be local variables, since this is a long
1215
search = self._file_ids_altered_regex.search
1216
unescape = _unescape_xml
1217
setdefault = result.setdefault
1218
for line, version_id in line_iterator:
1219
match = search(line)
1222
# One call to match.group() returning multiple items is quite a
1223
# bit faster than 2 calls to match.group() each returning 1
1224
file_id, revision_id = match.group('file_id', 'revision_id')
1226
# Inlining the cache lookups helps a lot when you make 170,000
1227
# lines and 350k ids, versus 8.4 unique ids.
1228
# Using a cache helps in 2 ways:
1229
# 1) Avoids unnecessary decoding calls
1230
# 2) Re-uses cached strings, which helps in future set and
1232
# (2) is enough that removing encoding entirely along with
1233
# the cache (so we are using plain strings) results in no
1234
# performance improvement.
1236
revision_id = unescape_revid_cache[revision_id]
1238
unescaped = unescape(revision_id)
1239
unescape_revid_cache[revision_id] = unescaped
1240
revision_id = unescaped
1242
# Note that unconditionally unescaping means that we deserialise
1243
# every fileid, which for general 'pull' is not great, but we don't
1244
# really want to have some many fulltexts that this matters anyway.
1247
file_id = unescape_fileid_cache[file_id]
1249
unescaped = unescape(file_id)
1250
unescape_fileid_cache[file_id] = unescaped
1253
key = (file_id, revision_id)
1254
setdefault(key, False)
1255
if revision_id == version_id:
1259
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1261
"""Helper routine for fileids_altered_by_revision_ids.
1263
This performs the translation of xml lines to revision ids.
1265
:param line_iterator: An iterator of lines, origin_version_id
1266
:param revision_ids: The revision ids to filter for. This should be a
1267
set or other type which supports efficient __contains__ lookups, as
1268
the revision id from each parsed line will be looked up in the
1269
revision_ids filter.
1270
:return: a dictionary mapping altered file-ids to an iterable of
1271
revision_ids. Each altered file-ids has the exact revision_ids that
1272
altered it listed explicitly.
1275
setdefault = result.setdefault
1276
for file_id, revision_id in \
1277
self._find_text_key_references_from_xml_inventory_lines(
1278
line_iterator).iterkeys():
1279
# once data is all ensured-consistent; then this is
1280
# if revision_id == version_id
1281
if revision_id in revision_ids:
1282
setdefault(file_id, set()).add(revision_id)
1285
def fileids_altered_by_revision_ids(self, revision_ids, _inv_weave=None):
1286
"""Find the file ids and versions affected by revisions.
1288
:param revisions: an iterable containing revision ids.
1289
:param _inv_weave: The inventory weave from this repository or None.
1290
If None, the inventory weave will be opened automatically.
1291
:return: a dictionary mapping altered file-ids to an iterable of
1292
revision_ids. Each altered file-ids has the exact revision_ids that
1293
altered it listed explicitly.
1295
selected_revision_ids = set(revision_ids)
1296
w = _inv_weave or self.get_inventory_weave()
1297
pb = ui.ui_factory.nested_progress_bar()
1299
return self._find_file_ids_from_xml_inventory_lines(
1300
w.iter_lines_added_or_present_in_versions(
1301
selected_revision_ids, pb=pb),
1302
selected_revision_ids)
1306
def iter_files_bytes(self, desired_files):
1307
"""Iterate through file versions.
1309
Files will not necessarily be returned in the order they occur in
1310
desired_files. No specific order is guaranteed.
1312
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1313
value supplied by the caller as part of desired_files. It should
1314
uniquely identify the file version in the caller's context. (Examples:
1315
an index number or a TreeTransform trans_id.)
1317
bytes_iterator is an iterable of bytestrings for the file. The
1318
kind of iterable and length of the bytestrings are unspecified, but for
1319
this implementation, it is a list of lines produced by
1320
VersionedFile.get_lines().
1322
:param desired_files: a list of (file_id, revision_id, identifier)
1325
transaction = self.get_transaction()
1326
for file_id, revision_id, callable_data in desired_files:
1328
weave = self.weave_store.get_weave(file_id, transaction)
1329
except errors.NoSuchFile:
1330
raise errors.NoSuchIdInRepository(self, file_id)
1331
yield callable_data, weave.get_lines(revision_id)
1333
def _generate_text_key_index(self, text_key_references=None,
1335
"""Generate a new text key index for the repository.
1337
This is an expensive function that will take considerable time to run.
1339
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1340
list of parents, also text keys. When a given key has no parents,
1341
the parents list will be [NULL_REVISION].
1343
# All revisions, to find inventory parents.
1344
if ancestors is None:
1345
graph = self.get_graph()
1346
ancestors = graph.get_parent_map(self.all_revision_ids())
1347
if text_key_references is None:
1348
text_key_references = self.find_text_key_references()
1349
pb = ui.ui_factory.nested_progress_bar()
1351
return self._do_generate_text_key_index(ancestors,
1352
text_key_references, pb)
1356
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1357
"""Helper for _generate_text_key_index to avoid deep nesting."""
1358
revision_order = tsort.topo_sort(ancestors)
1359
invalid_keys = set()
1361
for revision_id in revision_order:
1362
revision_keys[revision_id] = set()
1363
text_count = len(text_key_references)
1364
# a cache of the text keys to allow reuse; costs a dict of all the
1365
# keys, but saves a 2-tuple for every child of a given key.
1367
for text_key, valid in text_key_references.iteritems():
1369
invalid_keys.add(text_key)
1371
revision_keys[text_key[1]].add(text_key)
1372
text_key_cache[text_key] = text_key
1373
del text_key_references
1375
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1376
NULL_REVISION = _mod_revision.NULL_REVISION
1377
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1378
# too small for large or very branchy trees. However, for 55K path
1379
# trees, it would be easy to use too much memory trivially. Ideally we
1380
# could gauge this by looking at available real memory etc, but this is
1381
# always a tricky proposition.
1382
inventory_cache = lru_cache.LRUCache(10)
1383
batch_size = 10 # should be ~150MB on a 55K path tree
1384
batch_count = len(revision_order) / batch_size + 1
1386
pb.update("Calculating text parents.", processed_texts, text_count)
1387
for offset in xrange(batch_count):
1388
to_query = revision_order[offset * batch_size:(offset + 1) *
1392
for rev_tree in self.revision_trees(to_query):
1393
revision_id = rev_tree.get_revision_id()
1394
parent_ids = ancestors[revision_id]
1395
for text_key in revision_keys[revision_id]:
1396
pb.update("Calculating text parents.", processed_texts)
1397
processed_texts += 1
1398
candidate_parents = []
1399
for parent_id in parent_ids:
1400
parent_text_key = (text_key[0], parent_id)
1402
check_parent = parent_text_key not in \
1403
revision_keys[parent_id]
1405
# the parent parent_id is a ghost:
1406
check_parent = False
1407
# truncate the derived graph against this ghost.
1408
parent_text_key = None
1410
# look at the parent commit details inventories to
1411
# determine possible candidates in the per file graph.
1414
inv = inventory_cache[parent_id]
1416
inv = self.revision_tree(parent_id).inventory
1417
inventory_cache[parent_id] = inv
1418
parent_entry = inv._byid.get(text_key[0], None)
1419
if parent_entry is not None:
1421
text_key[0], parent_entry.revision)
1423
parent_text_key = None
1424
if parent_text_key is not None:
1425
candidate_parents.append(
1426
text_key_cache[parent_text_key])
1427
parent_heads = text_graph.heads(candidate_parents)
1428
new_parents = list(parent_heads)
1429
new_parents.sort(key=lambda x:candidate_parents.index(x))
1430
if new_parents == []:
1431
new_parents = [NULL_REVISION]
1432
text_index[text_key] = new_parents
1434
for text_key in invalid_keys:
1435
text_index[text_key] = [NULL_REVISION]
1438
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1439
"""Get an iterable listing the keys of all the data introduced by a set
1442
The keys will be ordered so that the corresponding items can be safely
1443
fetched and inserted in that order.
1445
:returns: An iterable producing tuples of (knit-kind, file-id,
1446
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1447
'revisions'. file-id is None unless knit-kind is 'file'.
1449
# XXX: it's a bit weird to control the inventory weave caching in this
1450
# generator. Ideally the caching would be done in fetch.py I think. Or
1451
# maybe this generator should explicitly have the contract that it
1452
# should not be iterated until the previously yielded item has been
1454
inv_w = self.get_inventory_weave()
1456
# file ids that changed
1457
file_ids = self.fileids_altered_by_revision_ids(revision_ids, inv_w)
1459
num_file_ids = len(file_ids)
1460
for file_id, altered_versions in file_ids.iteritems():
1461
if _files_pb is not None:
1462
_files_pb.update("fetch texts", count, num_file_ids)
1464
yield ("file", file_id, altered_versions)
1465
# We're done with the files_pb. Note that it finished by the caller,
1466
# just as it was created by the caller.
1470
yield ("inventory", None, revision_ids)
1473
revisions_with_signatures = set()
1474
for rev_id in revision_ids:
1476
self.get_signature_text(rev_id)
1477
except errors.NoSuchRevision:
1481
revisions_with_signatures.add(rev_id)
1482
yield ("signatures", None, revisions_with_signatures)
1485
yield ("revisions", None, revision_ids)
1488
def get_inventory_weave(self):
1489
return self.control_weaves.get_weave('inventory',
1490
self.get_transaction())
1493
def get_inventory(self, revision_id):
1494
"""Get Inventory object by revision id."""
1495
return self.iter_inventories([revision_id]).next()
1497
def iter_inventories(self, revision_ids):
1498
"""Get many inventories by revision_ids.
1500
This will buffer some or all of the texts used in constructing the
1501
inventories in memory, but will only parse a single inventory at a
1504
:return: An iterator of inventories.
1506
if ((None in revision_ids)
1507
or (_mod_revision.NULL_REVISION in revision_ids)):
1508
raise ValueError('cannot get null revision inventory')
1509
return self._iter_inventories(revision_ids)
1511
def _iter_inventories(self, revision_ids):
1512
"""single-document based inventory iteration."""
1513
texts = self.get_inventory_weave().get_texts(revision_ids)
1514
for text, revision_id in zip(texts, revision_ids):
1515
yield self.deserialise_inventory(revision_id, text)
1517
def deserialise_inventory(self, revision_id, xml):
1518
"""Transform the xml into an inventory object.
1520
:param revision_id: The expected revision id of the inventory.
1521
:param xml: A serialised inventory.
1523
result = self._serializer.read_inventory_from_string(xml, revision_id)
1524
if result.revision_id != revision_id:
1525
raise AssertionError('revision id mismatch %s != %s' % (
1526
result.revision_id, revision_id))
1529
def serialise_inventory(self, inv):
1530
return self._serializer.write_inventory_to_string(inv)
1532
def _serialise_inventory_to_lines(self, inv):
1533
return self._serializer.write_inventory_to_lines(inv)
1535
def get_serializer_format(self):
1536
return self._serializer.format_num
1539
def get_inventory_xml(self, revision_id):
1540
"""Get inventory XML as a file object."""
1542
iw = self.get_inventory_weave()
1543
return iw.get_text(revision_id)
1545
raise errors.HistoryMissing(self, 'inventory', revision_id)
1548
def get_inventory_sha1(self, revision_id):
1549
"""Return the sha1 hash of the inventory entry
1551
return self.get_revision(revision_id).inventory_sha1
1553
def iter_reverse_revision_history(self, revision_id):
1554
"""Iterate backwards through revision ids in the lefthand history
1556
:param revision_id: The revision id to start with. All its lefthand
1557
ancestors will be traversed.
1559
graph = self.get_graph()
1560
next_id = revision_id
1562
if next_id in (None, _mod_revision.NULL_REVISION):
1565
# Note: The following line may raise KeyError in the event of
1566
# truncated history. We decided not to have a try:except:raise
1567
# RevisionNotPresent here until we see a use for it, because of the
1568
# cost in an inner loop that is by its very nature O(history).
1569
# Robert Collins 20080326
1570
parents = graph.get_parent_map([next_id])[next_id]
1571
if len(parents) == 0:
1574
next_id = parents[0]
1577
def get_revision_inventory(self, revision_id):
1578
"""Return inventory of a past revision."""
1579
# TODO: Unify this with get_inventory()
1580
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1581
# must be the same as its revision, so this is trivial.
1582
if revision_id is None:
1583
# This does not make sense: if there is no revision,
1584
# then it is the current tree inventory surely ?!
1585
# and thus get_root_id() is something that looks at the last
1586
# commit on the branch, and the get_root_id is an inventory check.
1587
raise NotImplementedError
1588
# return Inventory(self.get_root_id())
1590
return self.get_inventory(revision_id)
1593
def is_shared(self):
1594
"""Return True if this repository is flagged as a shared repository."""
1595
raise NotImplementedError(self.is_shared)
1598
def reconcile(self, other=None, thorough=False):
1599
"""Reconcile this repository."""
1600
from bzrlib.reconcile import RepoReconciler
1601
reconciler = RepoReconciler(self, thorough=thorough)
1602
reconciler.reconcile()
1605
def _refresh_data(self):
1606
"""Helper called from lock_* to ensure coherency with disk.
1608
The default implementation does nothing; it is however possible
1609
for repositories to maintain loaded indices across multiple locks
1610
by checking inside their implementation of this method to see
1611
whether their indices are still valid. This depends of course on
1612
the disk format being validatable in this manner.
1616
def revision_tree(self, revision_id):
1617
"""Return Tree for a revision on this branch.
1619
`revision_id` may be None for the empty tree revision.
1621
# TODO: refactor this to use an existing revision object
1622
# so we don't need to read it in twice.
1623
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1624
return RevisionTree(self, Inventory(root_id=None),
1625
_mod_revision.NULL_REVISION)
1627
inv = self.get_revision_inventory(revision_id)
1628
return RevisionTree(self, inv, revision_id)
1630
def revision_trees(self, revision_ids):
1631
"""Return Tree for a revision on this branch.
1633
`revision_id` may not be None or 'null:'"""
1634
inventories = self.iter_inventories(revision_ids)
1635
for inv in inventories:
1636
yield RevisionTree(self, inv, inv.revision_id)
1639
def get_ancestry(self, revision_id, topo_sorted=True):
1640
"""Return a list of revision-ids integrated by a revision.
1642
The first element of the list is always None, indicating the origin
1643
revision. This might change when we have history horizons, or
1644
perhaps we should have a new API.
1646
This is topologically sorted.
1648
if _mod_revision.is_null(revision_id):
1650
if not self.has_revision(revision_id):
1651
raise errors.NoSuchRevision(self, revision_id)
1652
w = self.get_inventory_weave()
1653
candidates = w.get_ancestry(revision_id, topo_sorted)
1654
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1657
"""Compress the data within the repository.
1659
This operation only makes sense for some repository types. For other
1660
types it should be a no-op that just returns.
1662
This stub method does not require a lock, but subclasses should use
1663
@needs_write_lock as this is a long running call its reasonable to
1664
implicitly lock for the user.
1668
def print_file(self, file, revision_id):
1669
"""Print `file` to stdout.
1671
FIXME RBC 20060125 as John Meinel points out this is a bad api
1672
- it writes to stdout, it assumes that that is valid etc. Fix
1673
by creating a new more flexible convenience function.
1675
tree = self.revision_tree(revision_id)
1676
# use inventory as it was in that revision
1677
file_id = tree.inventory.path2id(file)
1679
# TODO: jam 20060427 Write a test for this code path
1680
# it had a bug in it, and was raising the wrong
1682
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1683
tree.print_file(file_id)
1685
def get_transaction(self):
1686
return self.control_files.get_transaction()
1688
@deprecated_method(symbol_versioning.one_five)
1689
def revision_parents(self, revision_id):
1690
return self.get_inventory_weave().parent_names(revision_id)
1692
@deprecated_method(symbol_versioning.one_one)
1693
def get_parents(self, revision_ids):
1694
"""See StackedParentsProvider.get_parents"""
1695
parent_map = self.get_parent_map(revision_ids)
1696
return [parent_map.get(r, None) for r in revision_ids]
1698
def get_parent_map(self, keys):
1699
"""See graph._StackedParentsProvider.get_parent_map"""
1701
for revision_id in keys:
1702
if revision_id is None:
1703
raise ValueError('get_parent_map(None) is not valid')
1704
if revision_id == _mod_revision.NULL_REVISION:
1705
parent_map[revision_id] = ()
1708
parent_id_list = self.get_revision(revision_id).parent_ids
1709
except errors.NoSuchRevision:
1712
if len(parent_id_list) == 0:
1713
parent_ids = (_mod_revision.NULL_REVISION,)
1715
parent_ids = tuple(parent_id_list)
1716
parent_map[revision_id] = parent_ids
1719
def _make_parents_provider(self):
1722
def get_graph(self, other_repository=None):
1723
"""Return the graph walker for this repository format"""
1724
parents_provider = self._make_parents_provider()
1725
if (other_repository is not None and
1726
not self.has_same_location(other_repository)):
1727
parents_provider = graph._StackedParentsProvider(
1728
[parents_provider, other_repository._make_parents_provider()])
1729
return graph.Graph(parents_provider)
1731
def _get_versioned_file_checker(self):
1732
"""Return an object suitable for checking versioned files."""
1733
return _VersionedFileChecker(self)
1735
def revision_ids_to_search_result(self, result_set):
1736
"""Convert a set of revision ids to a graph SearchResult."""
1737
result_parents = set()
1738
for parents in self.get_graph().get_parent_map(
1739
result_set).itervalues():
1740
result_parents.update(parents)
1741
included_keys = result_set.intersection(result_parents)
1742
start_keys = result_set.difference(included_keys)
1743
exclude_keys = result_parents.difference(result_set)
1744
result = graph.SearchResult(start_keys, exclude_keys,
1745
len(result_set), result_set)
1749
def set_make_working_trees(self, new_value):
1750
"""Set the policy flag for making working trees when creating branches.
1752
This only applies to branches that use this repository.
1754
The default is 'True'.
1755
:param new_value: True to restore the default, False to disable making
1758
raise NotImplementedError(self.set_make_working_trees)
1760
def make_working_trees(self):
1761
"""Returns the policy for making working trees on new branches."""
1762
raise NotImplementedError(self.make_working_trees)
1765
def sign_revision(self, revision_id, gpg_strategy):
1766
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1767
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1770
def has_signature_for_revision_id(self, revision_id):
1771
"""Query for a revision signature for revision_id in the repository."""
1772
return self._revision_store.has_signature(revision_id,
1773
self.get_transaction())
1776
def get_signature_text(self, revision_id):
1777
"""Return the text for a signature."""
1778
return self._revision_store.get_signature_text(revision_id,
1779
self.get_transaction())
1782
def check(self, revision_ids=None):
1783
"""Check consistency of all history of given revision_ids.
1785
Different repository implementations should override _check().
1787
:param revision_ids: A non-empty list of revision_ids whose ancestry
1788
will be checked. Typically the last revision_id of a branch.
1790
return self._check(revision_ids)
1792
def _check(self, revision_ids):
1793
result = check.Check(self)
1797
def _warn_if_deprecated(self):
1798
global _deprecation_warning_done
1799
if _deprecation_warning_done:
1801
_deprecation_warning_done = True
1802
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1803
% (self._format, self.bzrdir.transport.base))
1805
def supports_rich_root(self):
1806
return self._format.rich_root_data
1808
def _check_ascii_revisionid(self, revision_id, method):
1809
"""Private helper for ascii-only repositories."""
1810
# weave repositories refuse to store revisionids that are non-ascii.
1811
if revision_id is not None:
1812
# weaves require ascii revision ids.
1813
if isinstance(revision_id, unicode):
1815
revision_id.encode('ascii')
1816
except UnicodeEncodeError:
1817
raise errors.NonAsciiRevisionId(method, self)
1820
revision_id.decode('ascii')
1821
except UnicodeDecodeError:
1822
raise errors.NonAsciiRevisionId(method, self)
1824
def revision_graph_can_have_wrong_parents(self):
1825
"""Is it possible for this repository to have a revision graph with
1828
If True, then this repository must also implement
1829
_find_inconsistent_revision_parents so that check and reconcile can
1830
check for inconsistencies before proceeding with other checks that may
1831
depend on the revision index being consistent.
1833
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1836
# remove these delegates a while after bzr 0.15
1837
def __make_delegated(name, from_module):
1838
def _deprecated_repository_forwarder():
1839
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1840
% (name, from_module),
1843
m = __import__(from_module, globals(), locals(), [name])
1845
return getattr(m, name)
1846
except AttributeError:
1847
raise AttributeError('module %s has no name %s'
1849
globals()[name] = _deprecated_repository_forwarder
1852
'AllInOneRepository',
1853
'WeaveMetaDirRepository',
1854
'PreSplitOutRepositoryFormat',
1855
'RepositoryFormat4',
1856
'RepositoryFormat5',
1857
'RepositoryFormat6',
1858
'RepositoryFormat7',
1860
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1864
'RepositoryFormatKnit',
1865
'RepositoryFormatKnit1',
1867
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1870
def install_revision(repository, rev, revision_tree):
1871
"""Install all revision data into a repository."""
1872
install_revisions(repository, [(rev, revision_tree, None)])
1875
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1876
"""Install all revision data into a repository.
1878
Accepts an iterable of revision, tree, signature tuples. The signature
1881
repository.start_write_group()
1883
for n, (revision, revision_tree, signature) in enumerate(iterable):
1884
_install_revision(repository, revision, revision_tree, signature)
1886
pb.update('Transferring revisions', n + 1, num_revisions)
1888
repository.abort_write_group()
1891
repository.commit_write_group()
1894
def _install_revision(repository, rev, revision_tree, signature):
1895
"""Install all revision data into a repository."""
1896
present_parents = []
1898
for p_id in rev.parent_ids:
1899
if repository.has_revision(p_id):
1900
present_parents.append(p_id)
1901
parent_trees[p_id] = repository.revision_tree(p_id)
1903
parent_trees[p_id] = repository.revision_tree(None)
1905
inv = revision_tree.inventory
1906
entries = inv.iter_entries()
1907
# backwards compatibility hack: skip the root id.
1908
if not repository.supports_rich_root():
1909
path, root = entries.next()
1910
if root.revision != rev.revision_id:
1911
raise errors.IncompatibleRevision(repr(repository))
1912
# Add the texts that are not already present
1913
for path, ie in entries:
1914
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1915
repository.get_transaction())
1916
if ie.revision not in w:
1918
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1919
# with InventoryEntry.find_previous_heads(). if it is, then there
1920
# is a latent bug here where the parents may have ancestors of each
1922
for revision, tree in parent_trees.iteritems():
1923
if ie.file_id not in tree:
1925
parent_id = tree.inventory[ie.file_id].revision
1926
if parent_id in text_parents:
1928
text_parents.append(parent_id)
1930
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1931
repository.get_transaction())
1932
lines = revision_tree.get_file(ie.file_id).readlines()
1933
vfile.add_lines(rev.revision_id, text_parents, lines)
1935
# install the inventory
1936
repository.add_inventory(rev.revision_id, inv, present_parents)
1937
except errors.RevisionAlreadyPresent:
1939
if signature is not None:
1940
repository.add_signature_text(rev.revision_id, signature)
1941
repository.add_revision(rev.revision_id, rev, inv)
1944
class MetaDirRepository(Repository):
1945
"""Repositories in the new meta-dir layout.
1947
:ivar _transport: Transport for access to repository control files,
1948
typically pointing to .bzr/repository.
1951
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1952
super(MetaDirRepository, self).__init__(_format,
1958
self._transport = control_files._transport
1961
def is_shared(self):
1962
"""Return True if this repository is flagged as a shared repository."""
1963
return self._transport.has('shared-storage')
1966
def set_make_working_trees(self, new_value):
1967
"""Set the policy flag for making working trees when creating branches.
1969
This only applies to branches that use this repository.
1971
The default is 'True'.
1972
:param new_value: True to restore the default, False to disable making
1977
self._transport.delete('no-working-trees')
1978
except errors.NoSuchFile:
1981
self._transport.put_bytes('no-working-trees', '',
1982
mode=self.control_files._file_mode)
1984
def make_working_trees(self):
1985
"""Returns the policy for making working trees on new branches."""
1986
return not self._transport.has('no-working-trees')
1989
class MetaDirVersionedFileRepository(MetaDirRepository):
1990
"""Repositories in a meta-dir, that work via versioned file objects."""
1992
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1993
super(MetaDirVersionedFileRepository, self).__init__(_format, a_bzrdir,
1994
control_files, _revision_store, control_store, text_store)
1995
_revision_store.get_scope = self.get_transaction
1996
control_store.get_scope = self.get_transaction
1997
text_store.get_scope = self.get_transaction
2000
class RepositoryFormatRegistry(registry.Registry):
2001
"""Registry of RepositoryFormats."""
2003
def get(self, format_string):
2004
r = registry.Registry.get(self, format_string)
2010
format_registry = RepositoryFormatRegistry()
2011
"""Registry of formats, indexed by their identifying format string.
2013
This can contain either format instances themselves, or classes/factories that
2014
can be called to obtain one.
2018
#####################################################################
2019
# Repository Formats
2021
class RepositoryFormat(object):
2022
"""A repository format.
2024
Formats provide three things:
2025
* An initialization routine to construct repository data on disk.
2026
* a format string which is used when the BzrDir supports versioned
2028
* an open routine which returns a Repository instance.
2030
There is one and only one Format subclass for each on-disk format. But
2031
there can be one Repository subclass that is used for several different
2032
formats. The _format attribute on a Repository instance can be used to
2033
determine the disk format.
2035
Formats are placed in an dict by their format string for reference
2036
during opening. These should be subclasses of RepositoryFormat
2039
Once a format is deprecated, just deprecate the initialize and open
2040
methods on the format class. Do not deprecate the object, as the
2041
object will be created every system load.
2043
Common instance attributes:
2044
_matchingbzrdir - the bzrdir format that the repository format was
2045
originally written to work with. This can be used if manually
2046
constructing a bzrdir and repository, or more commonly for test suite
2050
# Set to True or False in derived classes. True indicates that the format
2051
# supports ghosts gracefully.
2052
supports_ghosts = None
2053
# Can this repository be given external locations to lookup additional
2054
# data. Set to True or False in derived classes.
2055
supports_external_lookups = None
2058
return "<%s>" % self.__class__.__name__
2060
def __eq__(self, other):
2061
# format objects are generally stateless
2062
return isinstance(other, self.__class__)
2064
def __ne__(self, other):
2065
return not self == other
2068
def find_format(klass, a_bzrdir):
2069
"""Return the format for the repository object in a_bzrdir.
2071
This is used by bzr native formats that have a "format" file in
2072
the repository. Other methods may be used by different types of
2076
transport = a_bzrdir.get_repository_transport(None)
2077
format_string = transport.get("format").read()
2078
return format_registry.get(format_string)
2079
except errors.NoSuchFile:
2080
raise errors.NoRepositoryPresent(a_bzrdir)
2082
raise errors.UnknownFormatError(format=format_string,
2086
def register_format(klass, format):
2087
format_registry.register(format.get_format_string(), format)
2090
def unregister_format(klass, format):
2091
format_registry.remove(format.get_format_string())
2094
def get_default_format(klass):
2095
"""Return the current default format."""
2096
from bzrlib import bzrdir
2097
return bzrdir.format_registry.make_bzrdir('default').repository_format
2099
def _get_control_store(self, repo_transport, control_files):
2100
"""Return the control store for this repository."""
2101
raise NotImplementedError(self._get_control_store)
2103
def get_format_string(self):
2104
"""Return the ASCII format string that identifies this format.
2106
Note that in pre format ?? repositories the format string is
2107
not permitted nor written to disk.
2109
raise NotImplementedError(self.get_format_string)
2111
def get_format_description(self):
2112
"""Return the short description for this format."""
2113
raise NotImplementedError(self.get_format_description)
2115
def _get_revision_store(self, repo_transport, control_files):
2116
"""Return the revision store object for this a_bzrdir."""
2117
raise NotImplementedError(self._get_revision_store)
2119
def _get_text_rev_store(self,
2126
"""Common logic for getting a revision store for a repository.
2128
see self._get_revision_store for the subclass-overridable method to
2129
get the store for a repository.
2131
from bzrlib.store.revision.text import TextRevisionStore
2132
dir_mode = control_files._dir_mode
2133
file_mode = control_files._file_mode
2134
text_store = TextStore(transport.clone(name),
2136
compressed=compressed,
2138
file_mode=file_mode)
2139
_revision_store = TextRevisionStore(text_store, serializer)
2140
return _revision_store
2142
# TODO: this shouldn't be in the base class, it's specific to things that
2143
# use weaves or knits -- mbp 20070207
2144
def _get_versioned_file_store(self,
2149
versionedfile_class=None,
2150
versionedfile_kwargs={},
2152
if versionedfile_class is None:
2153
versionedfile_class = self._versionedfile_class
2154
weave_transport = control_files._transport.clone(name)
2155
dir_mode = control_files._dir_mode
2156
file_mode = control_files._file_mode
2157
return VersionedFileStore(weave_transport, prefixed=prefixed,
2159
file_mode=file_mode,
2160
versionedfile_class=versionedfile_class,
2161
versionedfile_kwargs=versionedfile_kwargs,
2164
def initialize(self, a_bzrdir, shared=False):
2165
"""Initialize a repository of this format in a_bzrdir.
2167
:param a_bzrdir: The bzrdir to put the new repository in it.
2168
:param shared: The repository should be initialized as a sharable one.
2169
:returns: The new repository object.
2171
This may raise UninitializableFormat if shared repository are not
2172
compatible the a_bzrdir.
2174
raise NotImplementedError(self.initialize)
2176
def is_supported(self):
2177
"""Is this format supported?
2179
Supported formats must be initializable and openable.
2180
Unsupported formats may not support initialization or committing or
2181
some other features depending on the reason for not being supported.
2185
def check_conversion_target(self, target_format):
2186
raise NotImplementedError(self.check_conversion_target)
2188
def open(self, a_bzrdir, _found=False):
2189
"""Return an instance of this format for the bzrdir a_bzrdir.
2191
_found is a private parameter, do not use it.
2193
raise NotImplementedError(self.open)
2196
class MetaDirRepositoryFormat(RepositoryFormat):
2197
"""Common base class for the new repositories using the metadir layout."""
2199
rich_root_data = False
2200
supports_tree_reference = False
2201
supports_external_lookups = False
2202
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
2205
super(MetaDirRepositoryFormat, self).__init__()
2207
def _create_control_files(self, a_bzrdir):
2208
"""Create the required files and the initial control_files object."""
2209
# FIXME: RBC 20060125 don't peek under the covers
2210
# NB: no need to escape relative paths that are url safe.
2211
repository_transport = a_bzrdir.get_repository_transport(self)
2212
control_files = lockable_files.LockableFiles(repository_transport,
2213
'lock', lockdir.LockDir)
2214
control_files.create_lock()
2215
return control_files
2217
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
2218
"""Upload the initial blank content."""
2219
control_files = self._create_control_files(a_bzrdir)
2220
control_files.lock_write()
2221
transport = control_files._transport
2223
utf8_files += [('shared-storage', '')]
2225
transport.mkdir_multi(dirs, mode=a_bzrdir._get_dir_mode())
2226
for (filename, content_stream) in files:
2227
transport.put_file(filename, content_stream,
2228
mode=a_bzrdir._get_file_mode())
2229
for (filename, content_bytes) in utf8_files:
2230
transport.put_bytes_non_atomic(filename, content_bytes,
2231
mode=a_bzrdir._get_file_mode())
2233
control_files.unlock()
2236
# formats which have no format string are not discoverable
2237
# and not independently creatable, so are not registered. They're
2238
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
2239
# needed, it's constructed directly by the BzrDir. Non-native formats where
2240
# the repository is not separately opened are similar.
2242
format_registry.register_lazy(
2243
'Bazaar-NG Repository format 7',
2244
'bzrlib.repofmt.weaverepo',
2248
format_registry.register_lazy(
2249
'Bazaar-NG Knit Repository Format 1',
2250
'bzrlib.repofmt.knitrepo',
2251
'RepositoryFormatKnit1',
2254
format_registry.register_lazy(
2255
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
2256
'bzrlib.repofmt.knitrepo',
2257
'RepositoryFormatKnit3',
2260
format_registry.register_lazy(
2261
'Bazaar Knit Repository Format 4 (bzr 1.0)\n',
2262
'bzrlib.repofmt.knitrepo',
2263
'RepositoryFormatKnit4',
2266
# Pack-based formats. There is one format for pre-subtrees, and one for
2267
# post-subtrees to allow ease of testing.
2268
# NOTE: These are experimental in 0.92. Stable in 1.0 and above
2269
format_registry.register_lazy(
2270
'Bazaar pack repository format 1 (needs bzr 0.92)\n',
2271
'bzrlib.repofmt.pack_repo',
2272
'RepositoryFormatKnitPack1',
2274
format_registry.register_lazy(
2275
'Bazaar pack repository format 1 with subtree support (needs bzr 0.92)\n',
2276
'bzrlib.repofmt.pack_repo',
2277
'RepositoryFormatKnitPack3',
2279
format_registry.register_lazy(
2280
'Bazaar pack repository format 1 with rich root (needs bzr 1.0)\n',
2281
'bzrlib.repofmt.pack_repo',
2282
'RepositoryFormatKnitPack4',
2284
# Development formats.
2286
# development 0 - stub to introduce development versioning scheme.
2287
format_registry.register_lazy(
2288
"Bazaar development format 0 (needs bzr.dev from before 1.3)\n",
2289
'bzrlib.repofmt.pack_repo',
2290
'RepositoryFormatPackDevelopment0',
2292
format_registry.register_lazy(
2293
("Bazaar development format 0 with subtree support "
2294
"(needs bzr.dev from before 1.3)\n"),
2295
'bzrlib.repofmt.pack_repo',
2296
'RepositoryFormatPackDevelopment0Subtree',
2298
# 1.3->1.4 go below here
2301
class InterRepository(InterObject):
2302
"""This class represents operations taking place between two repositories.
2304
Its instances have methods like copy_content and fetch, and contain
2305
references to the source and target repositories these operations can be
2308
Often we will provide convenience methods on 'repository' which carry out
2309
operations with another repository - they will always forward to
2310
InterRepository.get(other).method_name(parameters).
2314
"""The available optimised InterRepository types."""
2316
def copy_content(self, revision_id=None):
2317
raise NotImplementedError(self.copy_content)
2319
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2320
"""Fetch the content required to construct revision_id.
2322
The content is copied from self.source to self.target.
2324
:param revision_id: if None all content is copied, if NULL_REVISION no
2326
:param pb: optional progress bar to use for progress reports. If not
2327
provided a default one will be created.
2329
Returns the copied revision count and the failed revisions in a tuple:
2332
raise NotImplementedError(self.fetch)
2334
def _walk_to_common_revisions(self, revision_ids):
2335
"""Walk out from revision_ids in source to revisions target has.
2337
:param revision_ids: The start point for the search.
2338
:return: A set of revision ids.
2340
target_graph = self.target.get_graph()
2341
revision_ids = frozenset(revision_ids)
2342
if set(target_graph.get_parent_map(revision_ids)) == revision_ids:
2343
return graph.SearchResult(revision_ids, set(), 0, set())
2344
missing_revs = set()
2345
source_graph = self.source.get_graph()
2346
# ensure we don't pay silly lookup costs.
2347
searcher = source_graph._make_breadth_first_searcher(revision_ids)
2348
null_set = frozenset([_mod_revision.NULL_REVISION])
2351
next_revs, ghosts = searcher.next_with_ghosts()
2352
except StopIteration:
2354
if revision_ids.intersection(ghosts):
2355
absent_ids = set(revision_ids.intersection(ghosts))
2356
# If all absent_ids are present in target, no error is needed.
2357
absent_ids.difference_update(
2358
set(target_graph.get_parent_map(absent_ids)))
2360
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2361
# we don't care about other ghosts as we can't fetch them and
2362
# haven't been asked to.
2363
next_revs = set(next_revs)
2364
# we always have NULL_REVISION present.
2365
have_revs = set(target_graph.get_parent_map(next_revs)).union(null_set)
2366
missing_revs.update(next_revs - have_revs)
2367
searcher.stop_searching_any(have_revs)
2368
return searcher.get_result()
2370
@deprecated_method(symbol_versioning.one_two)
2372
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2373
"""Return the revision ids that source has that target does not.
2375
These are returned in topological order.
2377
:param revision_id: only return revision ids included by this
2379
:param find_ghosts: If True find missing revisions in deep history
2380
rather than just finding the surface difference.
2382
return list(self.search_missing_revision_ids(
2383
revision_id, find_ghosts).get_keys())
2386
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2387
"""Return the revision ids that source has that target does not.
2389
:param revision_id: only return revision ids included by this
2391
:param find_ghosts: If True find missing revisions in deep history
2392
rather than just finding the surface difference.
2393
:return: A bzrlib.graph.SearchResult.
2395
# stop searching at found target revisions.
2396
if not find_ghosts and revision_id is not None:
2397
return self._walk_to_common_revisions([revision_id])
2398
# generic, possibly worst case, slow code path.
2399
target_ids = set(self.target.all_revision_ids())
2400
if revision_id is not None:
2401
source_ids = self.source.get_ancestry(revision_id)
2402
if source_ids[0] is not None:
2403
raise AssertionError()
2406
source_ids = self.source.all_revision_ids()
2407
result_set = set(source_ids).difference(target_ids)
2408
return self.source.revision_ids_to_search_result(result_set)
2411
def _same_model(source, target):
2412
"""True if source and target have the same data representation."""
2413
if source.supports_rich_root() != target.supports_rich_root():
2415
if source._serializer != target._serializer:
2420
class InterSameDataRepository(InterRepository):
2421
"""Code for converting between repositories that represent the same data.
2423
Data format and model must match for this to work.
2427
def _get_repo_format_to_test(self):
2428
"""Repository format for testing with.
2430
InterSameData can pull from subtree to subtree and from non-subtree to
2431
non-subtree, so we test this with the richest repository format.
2433
from bzrlib.repofmt import knitrepo
2434
return knitrepo.RepositoryFormatKnit3()
2437
def is_compatible(source, target):
2438
return InterRepository._same_model(source, target)
2441
def copy_content(self, revision_id=None):
2442
"""Make a complete copy of the content in self into destination.
2444
This copies both the repository's revision data, and configuration information
2445
such as the make_working_trees setting.
2447
This is a destructive operation! Do not use it on existing
2450
:param revision_id: Only copy the content needed to construct
2451
revision_id and its parents.
2454
self.target.set_make_working_trees(self.source.make_working_trees())
2455
except NotImplementedError:
2457
# but don't bother fetching if we have the needed data now.
2458
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2459
self.target.has_revision(revision_id)):
2461
self.target.fetch(self.source, revision_id=revision_id)
2464
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2465
"""See InterRepository.fetch()."""
2466
from bzrlib.fetch import GenericRepoFetcher
2467
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2468
self.source, self.source._format, self.target,
2469
self.target._format)
2470
f = GenericRepoFetcher(to_repository=self.target,
2471
from_repository=self.source,
2472
last_revision=revision_id,
2473
pb=pb, find_ghosts=find_ghosts)
2474
return f.count_copied, f.failed_revisions
2477
class InterWeaveRepo(InterSameDataRepository):
2478
"""Optimised code paths between Weave based repositories.
2480
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2481
implemented lazy inter-object optimisation.
2485
def _get_repo_format_to_test(self):
2486
from bzrlib.repofmt import weaverepo
2487
return weaverepo.RepositoryFormat7()
2490
def is_compatible(source, target):
2491
"""Be compatible with known Weave formats.
2493
We don't test for the stores being of specific types because that
2494
could lead to confusing results, and there is no need to be
2497
from bzrlib.repofmt.weaverepo import (
2503
return (isinstance(source._format, (RepositoryFormat5,
2505
RepositoryFormat7)) and
2506
isinstance(target._format, (RepositoryFormat5,
2508
RepositoryFormat7)))
2509
except AttributeError:
2513
def copy_content(self, revision_id=None):
2514
"""See InterRepository.copy_content()."""
2515
# weave specific optimised path:
2517
self.target.set_make_working_trees(self.source.make_working_trees())
2518
except (errors.RepositoryUpgradeRequired, NotImplemented):
2520
# FIXME do not peek!
2521
if self.source._transport.listable():
2522
pb = ui.ui_factory.nested_progress_bar()
2524
self.target.weave_store.copy_all_ids(
2525
self.source.weave_store,
2527
from_transaction=self.source.get_transaction(),
2528
to_transaction=self.target.get_transaction())
2529
pb.update('copying inventory', 0, 1)
2530
self.target.control_weaves.copy_multi(
2531
self.source.control_weaves, ['inventory'],
2532
from_transaction=self.source.get_transaction(),
2533
to_transaction=self.target.get_transaction())
2534
self.target._revision_store.text_store.copy_all_ids(
2535
self.source._revision_store.text_store,
2540
self.target.fetch(self.source, revision_id=revision_id)
2543
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2544
"""See InterRepository.fetch()."""
2545
from bzrlib.fetch import GenericRepoFetcher
2546
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2547
self.source, self.source._format, self.target, self.target._format)
2548
f = GenericRepoFetcher(to_repository=self.target,
2549
from_repository=self.source,
2550
last_revision=revision_id,
2551
pb=pb, find_ghosts=find_ghosts)
2552
return f.count_copied, f.failed_revisions
2555
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2556
"""See InterRepository.missing_revision_ids()."""
2557
# we want all revisions to satisfy revision_id in source.
2558
# but we don't want to stat every file here and there.
2559
# we want then, all revisions other needs to satisfy revision_id
2560
# checked, but not those that we have locally.
2561
# so the first thing is to get a subset of the revisions to
2562
# satisfy revision_id in source, and then eliminate those that
2563
# we do already have.
2564
# this is slow on high latency connection to self, but as as this
2565
# disk format scales terribly for push anyway due to rewriting
2566
# inventory.weave, this is considered acceptable.
2568
if revision_id is not None:
2569
source_ids = self.source.get_ancestry(revision_id)
2570
if source_ids[0] is not None:
2571
raise AssertionError()
2574
source_ids = self.source._all_possible_ids()
2575
source_ids_set = set(source_ids)
2576
# source_ids is the worst possible case we may need to pull.
2577
# now we want to filter source_ids against what we actually
2578
# have in target, but don't try to check for existence where we know
2579
# we do not have a revision as that would be pointless.
2580
target_ids = set(self.target._all_possible_ids())
2581
possibly_present_revisions = target_ids.intersection(source_ids_set)
2582
actually_present_revisions = set(
2583
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2584
required_revisions = source_ids_set.difference(actually_present_revisions)
2585
if revision_id is not None:
2586
# we used get_ancestry to determine source_ids then we are assured all
2587
# revisions referenced are present as they are installed in topological order.
2588
# and the tip revision was validated by get_ancestry.
2589
result_set = required_revisions
2591
# if we just grabbed the possibly available ids, then
2592
# we only have an estimate of whats available and need to validate
2593
# that against the revision records.
2595
self.source._eliminate_revisions_not_present(required_revisions))
2596
return self.source.revision_ids_to_search_result(result_set)
2599
class InterKnitRepo(InterSameDataRepository):
2600
"""Optimised code paths between Knit based repositories."""
2603
def _get_repo_format_to_test(self):
2604
from bzrlib.repofmt import knitrepo
2605
return knitrepo.RepositoryFormatKnit1()
2608
def is_compatible(source, target):
2609
"""Be compatible with known Knit formats.
2611
We don't test for the stores being of specific types because that
2612
could lead to confusing results, and there is no need to be
2615
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2617
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2618
isinstance(target._format, RepositoryFormatKnit))
2619
except AttributeError:
2621
return are_knits and InterRepository._same_model(source, target)
2624
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2625
"""See InterRepository.fetch()."""
2626
from bzrlib.fetch import KnitRepoFetcher
2627
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2628
self.source, self.source._format, self.target, self.target._format)
2629
f = KnitRepoFetcher(to_repository=self.target,
2630
from_repository=self.source,
2631
last_revision=revision_id,
2632
pb=pb, find_ghosts=find_ghosts)
2633
return f.count_copied, f.failed_revisions
2636
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2637
"""See InterRepository.missing_revision_ids()."""
2638
if revision_id is not None:
2639
source_ids = self.source.get_ancestry(revision_id)
2640
if source_ids[0] is not None:
2641
raise AssertionError()
2644
source_ids = self.source.all_revision_ids()
2645
source_ids_set = set(source_ids)
2646
# source_ids is the worst possible case we may need to pull.
2647
# now we want to filter source_ids against what we actually
2648
# have in target, but don't try to check for existence where we know
2649
# we do not have a revision as that would be pointless.
2650
target_ids = set(self.target.all_revision_ids())
2651
possibly_present_revisions = target_ids.intersection(source_ids_set)
2652
actually_present_revisions = set(
2653
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2654
required_revisions = source_ids_set.difference(actually_present_revisions)
2655
if revision_id is not None:
2656
# we used get_ancestry to determine source_ids then we are assured all
2657
# revisions referenced are present as they are installed in topological order.
2658
# and the tip revision was validated by get_ancestry.
2659
result_set = required_revisions
2661
# if we just grabbed the possibly available ids, then
2662
# we only have an estimate of whats available and need to validate
2663
# that against the revision records.
2665
self.source._eliminate_revisions_not_present(required_revisions))
2666
return self.source.revision_ids_to_search_result(result_set)
2669
class InterPackRepo(InterSameDataRepository):
2670
"""Optimised code paths between Pack based repositories."""
2673
def _get_repo_format_to_test(self):
2674
from bzrlib.repofmt import pack_repo
2675
return pack_repo.RepositoryFormatKnitPack1()
2678
def is_compatible(source, target):
2679
"""Be compatible with known Pack formats.
2681
We don't test for the stores being of specific types because that
2682
could lead to confusing results, and there is no need to be
2685
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2687
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2688
isinstance(target._format, RepositoryFormatPack))
2689
except AttributeError:
2691
return are_packs and InterRepository._same_model(source, target)
2694
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2695
"""See InterRepository.fetch()."""
2696
from bzrlib.repofmt.pack_repo import Packer
2697
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2698
self.source, self.source._format, self.target, self.target._format)
2699
self.count_copied = 0
2700
if revision_id is None:
2702
# everything to do - use pack logic
2703
# to fetch from all packs to one without
2704
# inventory parsing etc, IFF nothing to be copied is in the target.
2706
revision_ids = self.source.all_revision_ids()
2707
revision_keys = [(revid,) for revid in revision_ids]
2708
index = self.target._pack_collection.revision_index.combined_index
2709
present_revision_ids = set(item[1][0] for item in
2710
index.iter_entries(revision_keys))
2711
revision_ids = set(revision_ids) - present_revision_ids
2712
# implementing the TODO will involve:
2713
# - detecting when all of a pack is selected
2714
# - avoiding as much as possible pre-selection, so the
2715
# more-core routines such as create_pack_from_packs can filter in
2716
# a just-in-time fashion. (though having a HEADS list on a
2717
# repository might make this a lot easier, because we could
2718
# sensibly detect 'new revisions' without doing a full index scan.
2719
elif _mod_revision.is_null(revision_id):
2724
revision_ids = self.search_missing_revision_ids(revision_id,
2725
find_ghosts=find_ghosts).get_keys()
2726
except errors.NoSuchRevision:
2727
raise errors.InstallFailed([revision_id])
2728
if len(revision_ids) == 0:
2730
packs = self.source._pack_collection.all_packs()
2731
pack = Packer(self.target._pack_collection, packs, '.fetch',
2732
revision_ids).pack()
2733
if pack is not None:
2734
self.target._pack_collection._save_pack_names()
2735
# Trigger an autopack. This may duplicate effort as we've just done
2736
# a pack creation, but for now it is simpler to think about as
2737
# 'upload data, then repack if needed'.
2738
self.target._pack_collection.autopack()
2739
return (pack.get_revision_count(), [])
2744
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2745
"""See InterRepository.missing_revision_ids().
2747
:param find_ghosts: Find ghosts throughout the ancestry of
2750
if not find_ghosts and revision_id is not None:
2751
return self._walk_to_common_revisions([revision_id])
2752
elif revision_id is not None:
2753
source_ids = self.source.get_ancestry(revision_id)
2754
if source_ids[0] is not None:
2755
raise AssertionError()
2758
source_ids = self.source.all_revision_ids()
2759
# source_ids is the worst possible case we may need to pull.
2760
# now we want to filter source_ids against what we actually
2761
# have in target, but don't try to check for existence where we know
2762
# we do not have a revision as that would be pointless.
2763
target_ids = set(self.target.all_revision_ids())
2764
result_set = set(source_ids).difference(target_ids)
2765
return self.source.revision_ids_to_search_result(result_set)
2768
class InterModel1and2(InterRepository):
2771
def _get_repo_format_to_test(self):
2775
def is_compatible(source, target):
2776
if not source.supports_rich_root() and target.supports_rich_root():
2782
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2783
"""See InterRepository.fetch()."""
2784
from bzrlib.fetch import Model1toKnit2Fetcher
2785
f = Model1toKnit2Fetcher(to_repository=self.target,
2786
from_repository=self.source,
2787
last_revision=revision_id,
2788
pb=pb, find_ghosts=find_ghosts)
2789
return f.count_copied, f.failed_revisions
2792
def copy_content(self, revision_id=None):
2793
"""Make a complete copy of the content in self into destination.
2795
This is a destructive operation! Do not use it on existing
2798
:param revision_id: Only copy the content needed to construct
2799
revision_id and its parents.
2802
self.target.set_make_working_trees(self.source.make_working_trees())
2803
except NotImplementedError:
2805
# but don't bother fetching if we have the needed data now.
2806
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2807
self.target.has_revision(revision_id)):
2809
self.target.fetch(self.source, revision_id=revision_id)
2812
class InterKnit1and2(InterKnitRepo):
2815
def _get_repo_format_to_test(self):
2819
def is_compatible(source, target):
2820
"""Be compatible with Knit1 source and Knit3 target"""
2821
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2823
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2824
RepositoryFormatKnit3)
2825
from bzrlib.repofmt.pack_repo import (
2826
RepositoryFormatKnitPack1,
2827
RepositoryFormatKnitPack3,
2828
RepositoryFormatPackDevelopment0,
2829
RepositoryFormatPackDevelopment0Subtree,
2832
RepositoryFormatKnit1,
2833
RepositoryFormatKnitPack1,
2834
RepositoryFormatPackDevelopment0,
2837
RepositoryFormatKnit3,
2838
RepositoryFormatKnitPack3,
2839
RepositoryFormatPackDevelopment0Subtree,
2841
return (isinstance(source._format, nosubtrees) and
2842
isinstance(target._format, subtrees))
2843
except AttributeError:
2847
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2848
"""See InterRepository.fetch()."""
2849
from bzrlib.fetch import Knit1to2Fetcher
2850
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2851
self.source, self.source._format, self.target,
2852
self.target._format)
2853
f = Knit1to2Fetcher(to_repository=self.target,
2854
from_repository=self.source,
2855
last_revision=revision_id,
2856
pb=pb, find_ghosts=find_ghosts)
2857
return f.count_copied, f.failed_revisions
2860
class InterDifferingSerializer(InterKnitRepo):
2863
def _get_repo_format_to_test(self):
2867
def is_compatible(source, target):
2868
"""Be compatible with Knit2 source and Knit3 target"""
2869
if source.supports_rich_root() != target.supports_rich_root():
2871
# Ideally, we'd support fetching if the source had no tree references
2872
# even if it supported them...
2873
if (getattr(source, '_format.supports_tree_reference', False) and
2874
not getattr(target, '_format.supports_tree_reference', False)):
2879
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2880
"""See InterRepository.fetch()."""
2881
revision_ids = self.target.search_missing_revision_ids(self.source,
2882
revision_id, find_ghosts=find_ghosts).get_keys()
2883
revision_ids = tsort.topo_sort(
2884
self.source.get_graph().get_parent_map(revision_ids))
2885
def revisions_iterator():
2886
for current_revision_id in revision_ids:
2887
revision = self.source.get_revision(current_revision_id)
2888
tree = self.source.revision_tree(current_revision_id)
2890
signature = self.source.get_signature_text(
2891
current_revision_id)
2892
except errors.NoSuchRevision:
2894
yield revision, tree, signature
2896
my_pb = ui.ui_factory.nested_progress_bar()
2901
install_revisions(self.target, revisions_iterator(),
2902
len(revision_ids), pb)
2904
if my_pb is not None:
2906
return len(revision_ids), 0
2909
class InterRemoteToOther(InterRepository):
2911
def __init__(self, source, target):
2912
InterRepository.__init__(self, source, target)
2913
self._real_inter = None
2916
def is_compatible(source, target):
2917
if not isinstance(source, remote.RemoteRepository):
2919
# Is source's model compatible with target's model?
2920
source._ensure_real()
2921
real_source = source._real_repository
2922
if isinstance(real_source, remote.RemoteRepository):
2923
raise NotImplementedError(
2924
"We don't support remote repos backed by remote repos yet.")
2925
return InterRepository._same_model(real_source, target)
2928
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2929
"""See InterRepository.fetch()."""
2930
from bzrlib.fetch import RemoteToOtherFetcher
2931
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2932
self.source, self.target, self.target._format)
2933
# TODO: jam 20070210 This should be an assert, not a translate
2934
revision_id = osutils.safe_revision_id(revision_id)
2935
f = RemoteToOtherFetcher(to_repository=self.target,
2936
from_repository=self.source,
2937
last_revision=revision_id,
2938
pb=pb, find_ghosts=find_ghosts)
2939
return f.count_copied, f.failed_revisions
2942
def _get_repo_format_to_test(self):
2946
class InterOtherToRemote(InterRepository):
2948
def __init__(self, source, target):
2949
InterRepository.__init__(self, source, target)
2950
self._real_inter = None
2953
def is_compatible(source, target):
2954
if isinstance(target, remote.RemoteRepository):
2958
def _ensure_real_inter(self):
2959
if self._real_inter is None:
2960
self.target._ensure_real()
2961
real_target = self.target._real_repository
2962
self._real_inter = InterRepository.get(self.source, real_target)
2964
def copy_content(self, revision_id=None):
2965
self._ensure_real_inter()
2966
self._real_inter.copy_content(revision_id=revision_id)
2968
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2969
self._ensure_real_inter()
2970
self._real_inter.fetch(revision_id=revision_id, pb=pb,
2971
find_ghosts=find_ghosts)
2974
def _get_repo_format_to_test(self):
2978
InterRepository.register_optimiser(InterDifferingSerializer)
2979
InterRepository.register_optimiser(InterSameDataRepository)
2980
InterRepository.register_optimiser(InterWeaveRepo)
2981
InterRepository.register_optimiser(InterKnitRepo)
2982
InterRepository.register_optimiser(InterModel1and2)
2983
InterRepository.register_optimiser(InterKnit1and2)
2984
InterRepository.register_optimiser(InterPackRepo)
2985
InterRepository.register_optimiser(InterRemoteToOther)
2986
InterRepository.register_optimiser(InterOtherToRemote)
2989
class CopyConverter(object):
2990
"""A repository conversion tool which just performs a copy of the content.
2992
This is slow but quite reliable.
2995
def __init__(self, target_format):
2996
"""Create a CopyConverter.
2998
:param target_format: The format the resulting repository should be.
3000
self.target_format = target_format
3002
def convert(self, repo, pb):
3003
"""Perform the conversion of to_convert, giving feedback via pb.
3005
:param to_convert: The disk object to convert.
3006
:param pb: a progress bar to use for progress information.
3011
# this is only useful with metadir layouts - separated repo content.
3012
# trigger an assertion if not such
3013
repo._format.get_format_string()
3014
self.repo_dir = repo.bzrdir
3015
self.step('Moving repository to repository.backup')
3016
self.repo_dir.transport.move('repository', 'repository.backup')
3017
backup_transport = self.repo_dir.transport.clone('repository.backup')
3018
repo._format.check_conversion_target(self.target_format)
3019
self.source_repo = repo._format.open(self.repo_dir,
3021
_override_transport=backup_transport)
3022
self.step('Creating new repository')
3023
converted = self.target_format.initialize(self.repo_dir,
3024
self.source_repo.is_shared())
3025
converted.lock_write()
3027
self.step('Copying content into repository.')
3028
self.source_repo.copy_content_into(converted)
3031
self.step('Deleting old repository content.')
3032
self.repo_dir.transport.delete_tree('repository.backup')
3033
self.pb.note('repository converted')
3035
def step(self, message):
3036
"""Update the pb by a step."""
3038
self.pb.update(message, self.count, self.total)
3050
def _unescaper(match, _map=_unescape_map):
3051
code = match.group(1)
3055
if not code.startswith('#'):
3057
return unichr(int(code[1:])).encode('utf8')
3063
def _unescape_xml(data):
3064
"""Unescape predefined XML entities in a string of data."""
3066
if _unescape_re is None:
3067
_unescape_re = re.compile('\&([^;]*);')
3068
return _unescape_re.sub(_unescaper, data)
3071
class _VersionedFileChecker(object):
3073
def __init__(self, repository):
3074
self.repository = repository
3075
self.text_index = self.repository._generate_text_key_index()
3077
def calculate_file_version_parents(self, revision_id, file_id):
3078
"""Calculate the correct parents for a file version according to
3081
parent_keys = self.text_index[(file_id, revision_id)]
3082
if parent_keys == [_mod_revision.NULL_REVISION]:
3084
# strip the file_id, for the weave api
3085
return tuple([revision_id for file_id, revision_id in parent_keys])
3087
def check_file_version_parents(self, weave, file_id):
3088
"""Check the parents stored in a versioned file are correct.
3090
It also detects file versions that are not referenced by their
3091
corresponding revision's inventory.
3093
:returns: A tuple of (wrong_parents, dangling_file_versions).
3094
wrong_parents is a dict mapping {revision_id: (stored_parents,
3095
correct_parents)} for each revision_id where the stored parents
3096
are not correct. dangling_file_versions is a set of (file_id,
3097
revision_id) tuples for versions that are present in this versioned
3098
file, but not used by the corresponding inventory.
3101
unused_versions = set()
3102
versions = weave.versions()
3103
parent_map = weave.get_parent_map(versions)
3104
for num, revision_id in enumerate(versions):
3106
correct_parents = self.calculate_file_version_parents(
3107
revision_id, file_id)
3109
# The version is not part of the used keys.
3110
unused_versions.add(revision_id)
3113
knit_parents = tuple(parent_map[revision_id])
3114
except errors.RevisionNotPresent:
3116
if correct_parents != knit_parents:
3117
wrong_parents[revision_id] = (knit_parents, correct_parents)
3118
return wrong_parents, unused_versions
3121
def _old_get_graph(repository, revision_id):
3122
"""DO NOT USE. That is all. I'm serious."""
3123
graph = repository.get_graph()
3124
revision_graph = dict(((key, value) for key, value in
3125
graph.iter_ancestry([revision_id]) if value is not None))
3126
return _strip_NULL_ghosts(revision_graph)
3129
def _strip_NULL_ghosts(revision_graph):
3130
"""Also don't use this. more compatibility code for unmigrated clients."""
3131
# Filter ghosts, and null:
3132
if _mod_revision.NULL_REVISION in revision_graph:
3133
del revision_graph[_mod_revision.NULL_REVISION]
3134
for key, parents in revision_graph.items():
3135
revision_graph[key] = tuple(parent for parent in parents if parent
3137
return revision_graph
added
removed
bzrlib/repository.py
