1
# Copyright (C) 2005, 2006, 2007, 2008 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 (
63
from bzrlib.trace import mutter, mutter_callsite, note, warning
66
# Old formats display a warning, but only once
67
_deprecation_warning_done = False
70
class CommitBuilder(object):
71
"""Provides an interface to build up a commit.
73
This allows describing a tree to be committed without needing to
74
know the internals of the format of the repository.
77
# all clients should supply tree roots.
78
record_root_entry = True
79
# the default CommitBuilder does not manage trees whose root is versioned.
80
_versioned_root = False
82
def __init__(self, repository, parents, config, timestamp=None,
83
timezone=None, committer=None, revprops=None,
85
"""Initiate a CommitBuilder.
87
:param repository: Repository to commit to.
88
:param parents: Revision ids of the parents of the new revision.
89
:param config: Configuration to use.
90
:param timestamp: Optional timestamp recorded for commit.
91
:param timezone: Optional timezone for timestamp.
92
:param committer: Optional committer to set for commit.
93
:param revprops: Optional dictionary of revision properties.
94
:param revision_id: Optional revision id.
99
self._committer = self._config.username()
101
self._committer = committer
103
self.new_inventory = Inventory(None)
104
self._new_revision_id = revision_id
105
self.parents = parents
106
self.repository = repository
109
if revprops is not None:
110
self._revprops.update(revprops)
112
if timestamp is None:
113
timestamp = time.time()
114
# Restrict resolution to 1ms
115
self._timestamp = round(timestamp, 3)
118
self._timezone = osutils.local_time_offset()
120
self._timezone = int(timezone)
122
self._generate_revision_if_needed()
123
self.__heads = graph.HeadsCache(repository.get_graph()).heads
125
def commit(self, message):
126
"""Make the actual commit.
128
:return: The revision id of the recorded revision.
130
rev = _mod_revision.Revision(
131
timestamp=self._timestamp,
132
timezone=self._timezone,
133
committer=self._committer,
135
inventory_sha1=self.inv_sha1,
136
revision_id=self._new_revision_id,
137
properties=self._revprops)
138
rev.parent_ids = self.parents
139
self.repository.add_revision(self._new_revision_id, rev,
140
self.new_inventory, self._config)
141
self.repository.commit_write_group()
142
return self._new_revision_id
145
"""Abort the commit that is being built.
147
self.repository.abort_write_group()
149
def revision_tree(self):
150
"""Return the tree that was just committed.
152
After calling commit() this can be called to get a RevisionTree
153
representing the newly committed tree. This is preferred to
154
calling Repository.revision_tree() because that may require
155
deserializing the inventory, while we already have a copy in
158
return RevisionTree(self.repository, self.new_inventory,
159
self._new_revision_id)
161
def finish_inventory(self):
162
"""Tell the builder that the inventory is finished."""
163
if self.new_inventory.root is None:
164
raise AssertionError('Root entry should be supplied to'
165
' record_entry_contents, as of bzr 0.10.',
166
DeprecationWarning, stacklevel=2)
167
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
168
self.new_inventory.revision_id = self._new_revision_id
169
self.inv_sha1 = self.repository.add_inventory(
170
self._new_revision_id,
175
def _gen_revision_id(self):
176
"""Return new revision-id."""
177
return generate_ids.gen_revision_id(self._config.username(),
180
def _generate_revision_if_needed(self):
181
"""Create a revision id if None was supplied.
183
If the repository can not support user-specified revision ids
184
they should override this function and raise CannotSetRevisionId
185
if _new_revision_id is not None.
187
:raises: CannotSetRevisionId
189
if self._new_revision_id is None:
190
self._new_revision_id = self._gen_revision_id()
191
self.random_revid = True
193
self.random_revid = False
195
def _heads(self, file_id, revision_ids):
196
"""Calculate the graph heads for revision_ids in the graph of file_id.
198
This can use either a per-file graph or a global revision graph as we
199
have an identity relationship between the two graphs.
201
return self.__heads(revision_ids)
203
def _check_root(self, ie, parent_invs, tree):
204
"""Helper for record_entry_contents.
206
:param ie: An entry being added.
207
:param parent_invs: The inventories of the parent revisions of the
209
:param tree: The tree that is being committed.
211
# In this revision format, root entries have no knit or weave When
212
# serializing out to disk and back in root.revision is always
214
ie.revision = self._new_revision_id
216
def _get_delta(self, ie, basis_inv, path):
217
"""Get a delta against the basis inventory for ie."""
218
if ie.file_id not in basis_inv:
220
return (None, path, ie.file_id, ie)
221
elif ie != basis_inv[ie.file_id]:
223
# TODO: avoid tis id2path call.
224
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
229
def record_entry_contents(self, ie, parent_invs, path, tree,
231
"""Record the content of ie from tree into the commit if needed.
233
Side effect: sets ie.revision when unchanged
235
:param ie: An inventory entry present in the commit.
236
:param parent_invs: The inventories of the parent revisions of the
238
:param path: The path the entry is at in the tree.
239
:param tree: The tree which contains this entry and should be used to
241
:param content_summary: Summary data from the tree about the paths
242
content - stat, length, exec, sha/link target. This is only
243
accessed when the entry has a revision of None - that is when it is
244
a candidate to commit.
245
:return: A tuple (change_delta, version_recorded). change_delta is
246
an inventory_delta change for this entry against the basis tree of
247
the commit, or None if no change occured against the basis tree.
248
version_recorded is True if a new version of the entry has been
249
recorded. For instance, committing a merge where a file was only
250
changed on the other side will return (delta, False).
252
if self.new_inventory.root is None:
253
if ie.parent_id is not None:
254
raise errors.RootMissing()
255
self._check_root(ie, parent_invs, tree)
256
if ie.revision is None:
257
kind = content_summary[0]
259
# ie is carried over from a prior commit
261
# XXX: repository specific check for nested tree support goes here - if
262
# the repo doesn't want nested trees we skip it ?
263
if (kind == 'tree-reference' and
264
not self.repository._format.supports_tree_reference):
265
# mismatch between commit builder logic and repository:
266
# this needs the entry creation pushed down into the builder.
267
raise NotImplementedError('Missing repository subtree support.')
268
self.new_inventory.add(ie)
270
# TODO: slow, take it out of the inner loop.
272
basis_inv = parent_invs[0]
274
basis_inv = Inventory(root_id=None)
276
# ie.revision is always None if the InventoryEntry is considered
277
# for committing. We may record the previous parents revision if the
278
# content is actually unchanged against a sole head.
279
if ie.revision is not None:
280
if not self._versioned_root and path == '':
281
# repositories that do not version the root set the root's
282
# revision to the new commit even when no change occurs, and
283
# this masks when a change may have occurred against the basis,
284
# so calculate if one happened.
285
if ie.file_id in basis_inv:
286
delta = (basis_inv.id2path(ie.file_id), path,
290
delta = (None, path, ie.file_id, ie)
293
# we don't need to commit this, because the caller already
294
# determined that an existing revision of this file is
296
return None, (ie.revision == self._new_revision_id)
297
# XXX: Friction: parent_candidates should return a list not a dict
298
# so that we don't have to walk the inventories again.
299
parent_candiate_entries = ie.parent_candidates(parent_invs)
300
head_set = self._heads(ie.file_id, parent_candiate_entries.keys())
302
for inv in parent_invs:
303
if ie.file_id in inv:
304
old_rev = inv[ie.file_id].revision
305
if old_rev in head_set:
306
heads.append(inv[ie.file_id].revision)
307
head_set.remove(inv[ie.file_id].revision)
310
# now we check to see if we need to write a new record to the
312
# We write a new entry unless there is one head to the ancestors, and
313
# the kind-derived content is unchanged.
315
# Cheapest check first: no ancestors, or more the one head in the
316
# ancestors, we write a new node.
320
# There is a single head, look it up for comparison
321
parent_entry = parent_candiate_entries[heads[0]]
322
# if the non-content specific data has changed, we'll be writing a
324
if (parent_entry.parent_id != ie.parent_id or
325
parent_entry.name != ie.name):
327
# now we need to do content specific checks:
329
# if the kind changed the content obviously has
330
if kind != parent_entry.kind:
333
if content_summary[2] is None:
334
raise ValueError("Files must not have executable = None")
336
if (# if the file length changed we have to store:
337
parent_entry.text_size != content_summary[1] or
338
# if the exec bit has changed we have to store:
339
parent_entry.executable != content_summary[2]):
341
elif parent_entry.text_sha1 == content_summary[3]:
342
# all meta and content is unchanged (using a hash cache
343
# hit to check the sha)
344
ie.revision = parent_entry.revision
345
ie.text_size = parent_entry.text_size
346
ie.text_sha1 = parent_entry.text_sha1
347
ie.executable = parent_entry.executable
348
return self._get_delta(ie, basis_inv, path), False
350
# Either there is only a hash change(no hash cache entry,
351
# or same size content change), or there is no change on
353
# Provide the parent's hash to the store layer, so that the
354
# content is unchanged we will not store a new node.
355
nostore_sha = parent_entry.text_sha1
357
# We want to record a new node regardless of the presence or
358
# absence of a content change in the file.
360
ie.executable = content_summary[2]
361
lines = tree.get_file(ie.file_id, path).readlines()
363
ie.text_sha1, ie.text_size = self._add_text_to_weave(
364
ie.file_id, lines, heads, nostore_sha)
365
except errors.ExistingContent:
366
# Turns out that the file content was unchanged, and we were
367
# only going to store a new node if it was changed. Carry over
369
ie.revision = parent_entry.revision
370
ie.text_size = parent_entry.text_size
371
ie.text_sha1 = parent_entry.text_sha1
372
ie.executable = parent_entry.executable
373
return self._get_delta(ie, basis_inv, path), False
374
elif kind == 'directory':
376
# all data is meta here, nothing specific to directory, so
378
ie.revision = parent_entry.revision
379
return self._get_delta(ie, basis_inv, path), False
381
self._add_text_to_weave(ie.file_id, lines, heads, None)
382
elif kind == 'symlink':
383
current_link_target = content_summary[3]
385
# symlink target is not generic metadata, check if it has
387
if current_link_target != parent_entry.symlink_target:
390
# unchanged, carry over.
391
ie.revision = parent_entry.revision
392
ie.symlink_target = parent_entry.symlink_target
393
return self._get_delta(ie, basis_inv, path), False
394
ie.symlink_target = current_link_target
396
self._add_text_to_weave(ie.file_id, lines, heads, None)
397
elif kind == 'tree-reference':
399
if content_summary[3] != parent_entry.reference_revision:
402
# unchanged, carry over.
403
ie.reference_revision = parent_entry.reference_revision
404
ie.revision = parent_entry.revision
405
return self._get_delta(ie, basis_inv, path), False
406
ie.reference_revision = content_summary[3]
408
self._add_text_to_weave(ie.file_id, lines, heads, None)
410
raise NotImplementedError('unknown kind')
411
ie.revision = self._new_revision_id
412
return self._get_delta(ie, basis_inv, path), True
414
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
415
# Note: as we read the content directly from the tree, we know its not
416
# been turned into unicode or badly split - but a broken tree
417
# implementation could give us bad output from readlines() so this is
418
# not a guarantee of safety. What would be better is always checking
419
# the content during test suite execution. RBC 20070912
420
parent_keys = tuple((file_id, parent) for parent in parents)
421
return self.repository.texts.add_lines(
422
(file_id, self._new_revision_id), parent_keys, 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 some byte storage facilies (the revisions,
454
signatures, inventories and texts attributes) and a Transport, which
455
respectively provide byte storage and a means to access the (possibly
458
The byte storage facilities are addressed via tuples, which we refer to
459
as 'keys' throughout the code base. Revision_keys, inventory_keys and
460
signature_keys are all 1-tuples: (revision_id,). text_keys are two-tuples:
461
(file_id, revision_id). We use this interface because it allows low
462
friction with the underlying code that implements disk indices, network
463
encoding and other parts of bzrlib.
465
:ivar revisions: A bzrlib.versionedfile.VersionedFiles instance containing
466
the serialised revisions for the repository. This can be used to obtain
467
revision graph information or to access raw serialised revisions.
468
The result of trying to insert data into the repository via this store
469
is undefined: it should be considered read-only except for implementors
471
:ivar signatures: A bzrlib.versionedfile.VersionedFiles instance containing
472
the serialised signatures for the repository. This can be used to
473
obtain access to raw serialised signatures. The result of trying to
474
insert data into the repository via this store is undefined: it should
475
be considered read-only except for implementors of repositories.
476
:ivar inventories: A bzrlib.versionedfile.VersionedFiles instance containing
477
the serialised inventories for the repository. This can be used to
478
obtain unserialised inventories. The result of trying to insert data
479
into the repository via this store is undefined: it should be
480
considered read-only except for implementors of repositories.
481
:ivar texts: A bzrlib.versionedfile.VersionedFiles instance containing the
482
texts of files and directories for the repository. This can be used to
483
obtain file texts or file graphs. Note that Repository.iter_file_bytes
484
is usually a better interface for accessing file texts.
485
The result of trying to insert data into the repository via this store
486
is undefined: it should be considered read-only except for implementors
488
:ivar _transport: Transport for file access to repository, typically
489
pointing to .bzr/repository.
492
# What class to use for a CommitBuilder. Often its simpler to change this
493
# in a Repository class subclass rather than to override
494
# get_commit_builder.
495
_commit_builder_class = CommitBuilder
496
# The search regex used by xml based repositories to determine what things
497
# where changed in a single commit.
498
_file_ids_altered_regex = lazy_regex.lazy_compile(
499
r'file_id="(?P<file_id>[^"]+)"'
500
r'.* revision="(?P<revision_id>[^"]+)"'
503
def abort_write_group(self):
504
"""Commit the contents accrued within the current write group.
506
:seealso: start_write_group.
508
if self._write_group is not self.get_transaction():
509
# has an unlock or relock occured ?
510
raise errors.BzrError('mismatched lock context and write group.')
511
self._abort_write_group()
512
self._write_group = None
514
def _abort_write_group(self):
515
"""Template method for per-repository write group cleanup.
517
This is called during abort before the write group is considered to be
518
finished and should cleanup any internal state accrued during the write
519
group. There is no requirement that data handed to the repository be
520
*not* made available - this is not a rollback - but neither should any
521
attempt be made to ensure that data added is fully commited. Abort is
522
invoked when an error has occured so futher disk or network operations
523
may not be possible or may error and if possible should not be
527
def add_inventory(self, revision_id, inv, parents):
528
"""Add the inventory inv to the repository as revision_id.
530
:param parents: The revision ids of the parents that revision_id
531
is known to have and are in the repository already.
533
:returns: The validator(which is a sha1 digest, though what is sha'd is
534
repository format specific) of the serialized inventory.
536
if not self.is_in_write_group():
537
raise AssertionError("%r not in write group" % (self,))
538
_mod_revision.check_not_reserved_id(revision_id)
539
if not (inv.revision_id is None or inv.revision_id == revision_id):
540
raise AssertionError(
541
"Mismatch between inventory revision"
542
" id and insertion revid (%r, %r)"
543
% (inv.revision_id, revision_id))
545
raise AssertionError()
546
inv_lines = self._serialise_inventory_to_lines(inv)
547
return self._inventory_add_lines(revision_id, parents,
548
inv_lines, check_content=False)
550
def _inventory_add_lines(self, revision_id, parents, lines,
552
"""Store lines in inv_vf and return the sha1 of the inventory."""
553
parents = [(parent,) for parent in parents]
554
return self.inventories.add_lines((revision_id,), parents, lines,
555
check_content=check_content)[0]
557
def add_revision(self, revision_id, rev, inv=None, config=None):
558
"""Add rev to the revision store as revision_id.
560
:param revision_id: the revision id to use.
561
:param rev: The revision object.
562
:param inv: The inventory for the revision. if None, it will be looked
563
up in the inventory storer
564
:param config: If None no digital signature will be created.
565
If supplied its signature_needed method will be used
566
to determine if a signature should be made.
568
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
570
_mod_revision.check_not_reserved_id(revision_id)
571
if config is not None and config.signature_needed():
573
inv = self.get_inventory(revision_id)
574
plaintext = Testament(rev, inv).as_short_text()
575
self.store_revision_signature(
576
gpg.GPGStrategy(config), plaintext, revision_id)
577
# check inventory present
578
if not self.inventories.get_parent_map([(revision_id,)]):
580
raise errors.WeaveRevisionNotPresent(revision_id,
583
# yes, this is not suitable for adding with ghosts.
584
rev.inventory_sha1 = self.add_inventory(revision_id, inv,
587
rev.inventory_sha1 = self.inventories.get_sha1s([(revision_id,)])[0]
588
self._add_revision(rev)
590
def _add_revision(self, revision):
591
text = self._serializer.write_revision_to_string(revision)
592
key = (revision.revision_id,)
593
parents = tuple((parent,) for parent in revision.parent_ids)
594
self.revisions.add_lines(key, parents, osutils.split_lines(text))
596
def all_revision_ids(self):
597
"""Returns a list of all the revision ids in the repository.
599
This is deprecated because code should generally work on the graph
600
reachable from a particular revision, and ignore any other revisions
601
that might be present. There is no direct replacement method.
603
if 'evil' in debug.debug_flags:
604
mutter_callsite(2, "all_revision_ids is linear with history.")
605
return self._all_revision_ids()
607
def _all_revision_ids(self):
608
"""Returns a list of all the revision ids in the repository.
610
These are in as much topological order as the underlying store can
613
raise NotImplementedError(self._all_revision_ids)
615
def break_lock(self):
616
"""Break a lock if one is present from another instance.
618
Uses the ui factory to ask for confirmation if the lock may be from
621
self.control_files.break_lock()
624
def _eliminate_revisions_not_present(self, revision_ids):
625
"""Check every revision id in revision_ids to see if we have it.
627
Returns a set of the present revisions.
630
graph = self.get_graph()
631
parent_map = graph.get_parent_map(revision_ids)
632
# The old API returned a list, should this actually be a set?
633
return parent_map.keys()
636
def create(a_bzrdir):
637
"""Construct the current default format repository in a_bzrdir."""
638
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
640
def __init__(self, _format, a_bzrdir, control_files):
641
"""instantiate a Repository.
643
:param _format: The format of the repository on disk.
644
:param a_bzrdir: The BzrDir of the repository.
646
In the future we will have a single api for all stores for
647
getting file texts, inventories and revisions, then
648
this construct will accept instances of those things.
650
super(Repository, self).__init__()
651
self._format = _format
652
# the following are part of the public API for Repository:
653
self.bzrdir = a_bzrdir
654
self.control_files = control_files
655
self._transport = control_files._transport
656
self.base = self._transport.base
658
self._reconcile_does_inventory_gc = True
659
self._reconcile_fixes_text_parents = False
660
self._reconcile_backsup_inventory = True
661
# not right yet - should be more semantically clear ?
663
# TODO: make sure to construct the right store classes, etc, depending
664
# on whether escaping is required.
665
self._warn_if_deprecated()
666
self._write_group = None
669
return '%s(%r)' % (self.__class__.__name__,
672
def has_same_location(self, other):
673
"""Returns a boolean indicating if this repository is at the same
674
location as another repository.
676
This might return False even when two repository objects are accessing
677
the same physical repository via different URLs.
679
if self.__class__ is not other.__class__:
681
return (self._transport.base == other._transport.base)
683
def is_in_write_group(self):
684
"""Return True if there is an open write group.
686
:seealso: start_write_group.
688
return self._write_group is not None
691
return self.control_files.is_locked()
693
def is_write_locked(self):
694
"""Return True if this object is write locked."""
695
return self.is_locked() and self.control_files._lock_mode == 'w'
697
def lock_write(self, token=None):
698
"""Lock this repository for writing.
700
This causes caching within the repository obejct to start accumlating
701
data during reads, and allows a 'write_group' to be obtained. Write
702
groups must be used for actual data insertion.
704
:param token: if this is already locked, then lock_write will fail
705
unless the token matches the existing lock.
706
:returns: a token if this instance supports tokens, otherwise None.
707
:raises TokenLockingNotSupported: when a token is given but this
708
instance doesn't support using token locks.
709
:raises MismatchedToken: if the specified token doesn't match the token
710
of the existing lock.
711
:seealso: start_write_group.
713
A token should be passed in if you know that you have locked the object
714
some other way, and need to synchronise this object's state with that
717
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
719
result = self.control_files.lock_write(token=token)
724
self.control_files.lock_read()
727
def get_physical_lock_status(self):
728
return self.control_files.get_physical_lock_status()
730
def leave_lock_in_place(self):
731
"""Tell this repository not to release the physical lock when this
734
If lock_write doesn't return a token, then this method is not supported.
736
self.control_files.leave_in_place()
738
def dont_leave_lock_in_place(self):
739
"""Tell this repository to release the physical lock when this
740
object is unlocked, even if it didn't originally acquire it.
742
If lock_write doesn't return a token, then this method is not supported.
744
self.control_files.dont_leave_in_place()
747
def gather_stats(self, revid=None, committers=None):
748
"""Gather statistics from a revision id.
750
:param revid: The revision id to gather statistics from, if None, then
751
no revision specific statistics are gathered.
752
:param committers: Optional parameter controlling whether to grab
753
a count of committers from the revision specific statistics.
754
:return: A dictionary of statistics. Currently this contains:
755
committers: The number of committers if requested.
756
firstrev: A tuple with timestamp, timezone for the penultimate left
757
most ancestor of revid, if revid is not the NULL_REVISION.
758
latestrev: A tuple with timestamp, timezone for revid, if revid is
759
not the NULL_REVISION.
760
revisions: The total revision count in the repository.
761
size: An estimate disk size of the repository in bytes.
764
if revid and committers:
765
result['committers'] = 0
766
if revid and revid != _mod_revision.NULL_REVISION:
768
all_committers = set()
769
revisions = self.get_ancestry(revid)
770
# pop the leading None
772
first_revision = None
774
# ignore the revisions in the middle - just grab first and last
775
revisions = revisions[0], revisions[-1]
776
for revision in self.get_revisions(revisions):
777
if not first_revision:
778
first_revision = revision
780
all_committers.add(revision.committer)
781
last_revision = revision
783
result['committers'] = len(all_committers)
784
result['firstrev'] = (first_revision.timestamp,
785
first_revision.timezone)
786
result['latestrev'] = (last_revision.timestamp,
787
last_revision.timezone)
789
# now gather global repository information
790
# XXX: This is available for many repos regardless of listability.
791
if self.bzrdir.root_transport.listable():
792
# XXX: do we want to __define len__() ?
793
# Maybe the versionedfiles object should provide a different
794
# method to get the number of keys.
795
result['revisions'] = len(self.revisions.keys())
799
def find_branches(self, using=False):
800
"""Find branches underneath this repository.
802
This will include branches inside other branches.
804
:param using: If True, list only branches using this repository.
806
if using and not self.is_shared():
808
return [self.bzrdir.open_branch()]
809
except errors.NotBranchError:
811
class Evaluator(object):
814
self.first_call = True
816
def __call__(self, bzrdir):
817
# On the first call, the parameter is always the bzrdir
818
# containing the current repo.
819
if not self.first_call:
821
repository = bzrdir.open_repository()
822
except errors.NoRepositoryPresent:
825
return False, (None, repository)
826
self.first_call = False
828
value = (bzrdir.open_branch(), None)
829
except errors.NotBranchError:
834
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
835
self.bzrdir.root_transport, evaluate=Evaluator()):
836
if branch is not None:
837
branches.append(branch)
838
if not using and repository is not None:
839
branches.extend(repository.find_branches())
843
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
844
"""Return the revision ids that other has that this does not.
846
These are returned in topological order.
848
revision_id: only return revision ids included by revision_id.
850
return InterRepository.get(other, self).search_missing_revision_ids(
851
revision_id, find_ghosts)
853
@deprecated_method(one_two)
855
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
856
"""Return the revision ids that other has that this does not.
858
These are returned in topological order.
860
revision_id: only return revision ids included by revision_id.
862
keys = self.search_missing_revision_ids(
863
other, revision_id, find_ghosts).get_keys()
866
parents = other.get_graph().get_parent_map(keys)
869
return tsort.topo_sort(parents)
873
"""Open the repository rooted at base.
875
For instance, if the repository is at URL/.bzr/repository,
876
Repository.open(URL) -> a Repository instance.
878
control = bzrdir.BzrDir.open(base)
879
return control.open_repository()
881
def copy_content_into(self, destination, revision_id=None):
882
"""Make a complete copy of the content in self into destination.
884
This is a destructive operation! Do not use it on existing
887
return InterRepository.get(self, destination).copy_content(revision_id)
889
def commit_write_group(self):
890
"""Commit the contents accrued within the current write group.
892
:seealso: start_write_group.
894
if self._write_group is not self.get_transaction():
895
# has an unlock or relock occured ?
896
raise errors.BzrError('mismatched lock context %r and '
898
(self.get_transaction(), self._write_group))
899
self._commit_write_group()
900
self._write_group = None
902
def _commit_write_group(self):
903
"""Template method for per-repository write group cleanup.
905
This is called before the write group is considered to be
906
finished and should ensure that all data handed to the repository
907
for writing during the write group is safely committed (to the
908
extent possible considering file system caching etc).
911
def fetch(self, source, revision_id=None, pb=None, find_ghosts=False):
912
"""Fetch the content required to construct revision_id from source.
914
If revision_id is None all content is copied.
915
:param find_ghosts: Find and copy revisions in the source that are
916
ghosts in the target (and not reachable directly by walking out to
917
the first-present revision in target from revision_id).
919
# fast path same-url fetch operations
920
if self.has_same_location(source):
921
# check that last_revision is in 'from' and then return a
923
if (revision_id is not None and
924
not _mod_revision.is_null(revision_id)):
925
self.get_revision(revision_id)
927
inter = InterRepository.get(source, self)
929
return inter.fetch(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts)
930
except NotImplementedError:
931
raise errors.IncompatibleRepositories(source, self)
933
def create_bundle(self, target, base, fileobj, format=None):
934
return serializer.write_bundle(self, target, base, fileobj, format)
936
def get_commit_builder(self, branch, parents, config, timestamp=None,
937
timezone=None, committer=None, revprops=None,
939
"""Obtain a CommitBuilder for this repository.
941
:param branch: Branch to commit to.
942
:param parents: Revision ids of the parents of the new revision.
943
:param config: Configuration to use.
944
:param timestamp: Optional timestamp recorded for commit.
945
:param timezone: Optional timezone for timestamp.
946
:param committer: Optional committer to set for commit.
947
:param revprops: Optional dictionary of revision properties.
948
:param revision_id: Optional revision id.
950
result = self._commit_builder_class(self, parents, config,
951
timestamp, timezone, committer, revprops, revision_id)
952
self.start_write_group()
956
if (self.control_files._lock_count == 1 and
957
self.control_files._lock_mode == 'w'):
958
if self._write_group is not None:
959
self.abort_write_group()
960
self.control_files.unlock()
961
raise errors.BzrError(
962
'Must end write groups before releasing write locks.')
963
self.control_files.unlock()
966
def clone(self, a_bzrdir, revision_id=None):
967
"""Clone this repository into a_bzrdir using the current format.
969
Currently no check is made that the format of this repository and
970
the bzrdir format are compatible. FIXME RBC 20060201.
972
:return: The newly created destination repository.
974
# TODO: deprecate after 0.16; cloning this with all its settings is
975
# probably not very useful -- mbp 20070423
976
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
977
self.copy_content_into(dest_repo, revision_id)
980
def start_write_group(self):
981
"""Start a write group in the repository.
983
Write groups are used by repositories which do not have a 1:1 mapping
984
between file ids and backend store to manage the insertion of data from
985
both fetch and commit operations.
987
A write lock is required around the start_write_group/commit_write_group
988
for the support of lock-requiring repository formats.
990
One can only insert data into a repository inside a write group.
994
if not self.is_write_locked():
995
raise errors.NotWriteLocked(self)
996
if self._write_group:
997
raise errors.BzrError('already in a write group')
998
self._start_write_group()
999
# so we can detect unlock/relock - the write group is now entered.
1000
self._write_group = self.get_transaction()
1002
def _start_write_group(self):
1003
"""Template method for per-repository write group startup.
1005
This is called before the write group is considered to be
1010
def sprout(self, to_bzrdir, revision_id=None):
1011
"""Create a descendent repository for new development.
1013
Unlike clone, this does not copy the settings of the repository.
1015
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
1016
dest_repo.fetch(self, revision_id=revision_id)
1019
def _create_sprouting_repo(self, a_bzrdir, shared):
1020
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
1021
# use target default format.
1022
dest_repo = a_bzrdir.create_repository()
1024
# Most control formats need the repository to be specifically
1025
# created, but on some old all-in-one formats it's not needed
1027
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
1028
except errors.UninitializableFormat:
1029
dest_repo = a_bzrdir.open_repository()
1033
def has_revision(self, revision_id):
1034
"""True if this repository has a copy of the revision."""
1035
return revision_id in self.has_revisions((revision_id,))
1038
def has_revisions(self, revision_ids):
1039
"""Probe to find out the presence of multiple revisions.
1041
:param revision_ids: An iterable of revision_ids.
1042
:return: A set of the revision_ids that were present.
1044
parent_map = self.revisions.get_parent_map(
1045
[(rev_id,) for rev_id in revision_ids])
1047
if _mod_revision.NULL_REVISION in revision_ids:
1048
result.add(_mod_revision.NULL_REVISION)
1049
result.update([key[0] for key in parent_map])
1053
def get_revision(self, revision_id):
1054
"""Return the Revision object for a named revision."""
1055
return self.get_revisions([revision_id])[0]
1058
def get_revision_reconcile(self, revision_id):
1059
"""'reconcile' helper routine that allows access to a revision always.
1061
This variant of get_revision does not cross check the weave graph
1062
against the revision one as get_revision does: but it should only
1063
be used by reconcile, or reconcile-alike commands that are correcting
1064
or testing the revision graph.
1066
return self._get_revisions([revision_id])[0]
1069
def get_revisions(self, revision_ids):
1070
"""Get many revisions at once."""
1071
return self._get_revisions(revision_ids)
1074
def _get_revisions(self, revision_ids):
1075
"""Core work logic to get many revisions without sanity checks."""
1076
for rev_id in revision_ids:
1077
if not rev_id or not isinstance(rev_id, basestring):
1078
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1079
keys = [(key,) for key in revision_ids]
1080
stream = self.revisions.get_record_stream(keys, 'unordered', True)
1082
for record in stream:
1083
if record.storage_kind == 'absent':
1084
raise errors.NoSuchRevision(self, record.key[0])
1085
text = record.get_bytes_as('fulltext')
1086
rev = self._serializer.read_revision_from_string(text)
1087
revs[record.key[0]] = rev
1088
return [revs[revid] for revid in revision_ids]
1091
def get_revision_xml(self, revision_id):
1092
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1093
# would have already do it.
1094
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1095
rev = self.get_revision(revision_id)
1096
rev_tmp = StringIO()
1097
# the current serializer..
1098
self._serializer.write_revision(rev, rev_tmp)
1100
return rev_tmp.getvalue()
1102
def get_deltas_for_revisions(self, revisions):
1103
"""Produce a generator of revision deltas.
1105
Note that the input is a sequence of REVISIONS, not revision_ids.
1106
Trees will be held in memory until the generator exits.
1107
Each delta is relative to the revision's lefthand predecessor.
1109
required_trees = set()
1110
for revision in revisions:
1111
required_trees.add(revision.revision_id)
1112
required_trees.update(revision.parent_ids[:1])
1113
trees = dict((t.get_revision_id(), t) for
1114
t in self.revision_trees(required_trees))
1115
for revision in revisions:
1116
if not revision.parent_ids:
1117
old_tree = self.revision_tree(None)
1119
old_tree = trees[revision.parent_ids[0]]
1120
yield trees[revision.revision_id].changes_from(old_tree)
1123
def get_revision_delta(self, revision_id):
1124
"""Return the delta for one revision.
1126
The delta is relative to the left-hand predecessor of the
1129
r = self.get_revision(revision_id)
1130
return list(self.get_deltas_for_revisions([r]))[0]
1133
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1134
signature = gpg_strategy.sign(plaintext)
1135
self.add_signature_text(revision_id, signature)
1138
def add_signature_text(self, revision_id, signature):
1139
self.signatures.add_lines((revision_id,), (),
1140
osutils.split_lines(signature))
1142
def find_text_key_references(self):
1143
"""Find the text key references within the repository.
1145
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1146
revision_ids. Each altered file-ids has the exact revision_ids that
1147
altered it listed explicitly.
1148
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1149
to whether they were referred to by the inventory of the
1150
revision_id that they contain. The inventory texts from all present
1151
revision ids are assessed to generate this report.
1153
revision_keys = self.revisions.keys()
1154
w = self.inventories
1155
pb = ui.ui_factory.nested_progress_bar()
1157
return self._find_text_key_references_from_xml_inventory_lines(
1158
w.iter_lines_added_or_present_in_keys(revision_keys, pb=pb))
1162
def _find_text_key_references_from_xml_inventory_lines(self,
1164
"""Core routine for extracting references to texts from inventories.
1166
This performs the translation of xml lines to revision ids.
1168
:param line_iterator: An iterator of lines, origin_version_id
1169
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1170
to whether they were referred to by the inventory of the
1171
revision_id that they contain. Note that if that revision_id was
1172
not part of the line_iterator's output then False will be given -
1173
even though it may actually refer to that key.
1175
if not self._serializer.support_altered_by_hack:
1176
raise AssertionError(
1177
"_find_text_key_references_from_xml_inventory_lines only "
1178
"supported for branches which store inventory as unnested xml"
1179
", not on %r" % self)
1182
# this code needs to read every new line in every inventory for the
1183
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1184
# not present in one of those inventories is unnecessary but not
1185
# harmful because we are filtering by the revision id marker in the
1186
# inventory lines : we only select file ids altered in one of those
1187
# revisions. We don't need to see all lines in the inventory because
1188
# only those added in an inventory in rev X can contain a revision=X
1190
unescape_revid_cache = {}
1191
unescape_fileid_cache = {}
1193
# jam 20061218 In a big fetch, this handles hundreds of thousands
1194
# of lines, so it has had a lot of inlining and optimizing done.
1195
# Sorry that it is a little bit messy.
1196
# Move several functions to be local variables, since this is a long
1198
search = self._file_ids_altered_regex.search
1199
unescape = _unescape_xml
1200
setdefault = result.setdefault
1201
for line, line_key in line_iterator:
1202
match = search(line)
1205
# One call to match.group() returning multiple items is quite a
1206
# bit faster than 2 calls to match.group() each returning 1
1207
file_id, revision_id = match.group('file_id', 'revision_id')
1209
# Inlining the cache lookups helps a lot when you make 170,000
1210
# lines and 350k ids, versus 8.4 unique ids.
1211
# Using a cache helps in 2 ways:
1212
# 1) Avoids unnecessary decoding calls
1213
# 2) Re-uses cached strings, which helps in future set and
1215
# (2) is enough that removing encoding entirely along with
1216
# the cache (so we are using plain strings) results in no
1217
# performance improvement.
1219
revision_id = unescape_revid_cache[revision_id]
1221
unescaped = unescape(revision_id)
1222
unescape_revid_cache[revision_id] = unescaped
1223
revision_id = unescaped
1225
# Note that unconditionally unescaping means that we deserialise
1226
# every fileid, which for general 'pull' is not great, but we don't
1227
# really want to have some many fulltexts that this matters anyway.
1230
file_id = unescape_fileid_cache[file_id]
1232
unescaped = unescape(file_id)
1233
unescape_fileid_cache[file_id] = unescaped
1236
key = (file_id, revision_id)
1237
setdefault(key, False)
1238
if revision_id == line_key[-1]:
1242
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1244
"""Helper routine for fileids_altered_by_revision_ids.
1246
This performs the translation of xml lines to revision ids.
1248
:param line_iterator: An iterator of lines, origin_version_id
1249
:param revision_ids: The revision ids to filter for. This should be a
1250
set or other type which supports efficient __contains__ lookups, as
1251
the revision id from each parsed line will be looked up in the
1252
revision_ids filter.
1253
:return: a dictionary mapping altered file-ids to an iterable of
1254
revision_ids. Each altered file-ids has the exact revision_ids that
1255
altered it listed explicitly.
1258
setdefault = result.setdefault
1260
self._find_text_key_references_from_xml_inventory_lines(
1261
line_iterator).iterkeys():
1262
# once data is all ensured-consistent; then this is
1263
# if revision_id == version_id
1264
if key[-1:] in revision_ids:
1265
setdefault(key[0], set()).add(key[-1])
1268
def fileids_altered_by_revision_ids(self, revision_ids, _inv_weave=None):
1269
"""Find the file ids and versions affected by revisions.
1271
:param revisions: an iterable containing revision ids.
1272
:param _inv_weave: The inventory weave from this repository or None.
1273
If None, the inventory weave will be opened automatically.
1274
:return: a dictionary mapping altered file-ids to an iterable of
1275
revision_ids. Each altered file-ids has the exact revision_ids that
1276
altered it listed explicitly.
1278
selected_keys = set((revid,) for revid in revision_ids)
1279
w = _inv_weave or self.inventories
1280
pb = ui.ui_factory.nested_progress_bar()
1282
return self._find_file_ids_from_xml_inventory_lines(
1283
w.iter_lines_added_or_present_in_keys(
1284
selected_keys, pb=pb),
1289
def iter_files_bytes(self, desired_files):
1290
"""Iterate through file versions.
1292
Files will not necessarily be returned in the order they occur in
1293
desired_files. No specific order is guaranteed.
1295
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1296
value supplied by the caller as part of desired_files. It should
1297
uniquely identify the file version in the caller's context. (Examples:
1298
an index number or a TreeTransform trans_id.)
1300
bytes_iterator is an iterable of bytestrings for the file. The
1301
kind of iterable and length of the bytestrings are unspecified, but for
1302
this implementation, it is a list of bytes produced by
1303
VersionedFile.get_record_stream().
1305
:param desired_files: a list of (file_id, revision_id, identifier)
1308
transaction = self.get_transaction()
1310
for file_id, revision_id, callable_data in desired_files:
1311
text_keys[(file_id, revision_id)] = callable_data
1312
for record in self.texts.get_record_stream(text_keys, 'unordered', True):
1313
if record.storage_kind == 'absent':
1314
raise errors.RevisionNotPresent(record.key, self)
1315
yield text_keys[record.key], record.get_bytes_as('fulltext')
1317
def _generate_text_key_index(self, text_key_references=None,
1319
"""Generate a new text key index for the repository.
1321
This is an expensive function that will take considerable time to run.
1323
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1324
list of parents, also text keys. When a given key has no parents,
1325
the parents list will be [NULL_REVISION].
1327
# All revisions, to find inventory parents.
1328
if ancestors is None:
1329
graph = self.get_graph()
1330
ancestors = graph.get_parent_map(self.all_revision_ids())
1331
if text_key_references is None:
1332
text_key_references = self.find_text_key_references()
1333
pb = ui.ui_factory.nested_progress_bar()
1335
return self._do_generate_text_key_index(ancestors,
1336
text_key_references, pb)
1340
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1341
"""Helper for _generate_text_key_index to avoid deep nesting."""
1342
revision_order = tsort.topo_sort(ancestors)
1343
invalid_keys = set()
1345
for revision_id in revision_order:
1346
revision_keys[revision_id] = set()
1347
text_count = len(text_key_references)
1348
# a cache of the text keys to allow reuse; costs a dict of all the
1349
# keys, but saves a 2-tuple for every child of a given key.
1351
for text_key, valid in text_key_references.iteritems():
1353
invalid_keys.add(text_key)
1355
revision_keys[text_key[1]].add(text_key)
1356
text_key_cache[text_key] = text_key
1357
del text_key_references
1359
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1360
NULL_REVISION = _mod_revision.NULL_REVISION
1361
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1362
# too small for large or very branchy trees. However, for 55K path
1363
# trees, it would be easy to use too much memory trivially. Ideally we
1364
# could gauge this by looking at available real memory etc, but this is
1365
# always a tricky proposition.
1366
inventory_cache = lru_cache.LRUCache(10)
1367
batch_size = 10 # should be ~150MB on a 55K path tree
1368
batch_count = len(revision_order) / batch_size + 1
1370
pb.update("Calculating text parents.", processed_texts, text_count)
1371
for offset in xrange(batch_count):
1372
to_query = revision_order[offset * batch_size:(offset + 1) *
1376
for rev_tree in self.revision_trees(to_query):
1377
revision_id = rev_tree.get_revision_id()
1378
parent_ids = ancestors[revision_id]
1379
for text_key in revision_keys[revision_id]:
1380
pb.update("Calculating text parents.", processed_texts)
1381
processed_texts += 1
1382
candidate_parents = []
1383
for parent_id in parent_ids:
1384
parent_text_key = (text_key[0], parent_id)
1386
check_parent = parent_text_key not in \
1387
revision_keys[parent_id]
1389
# the parent parent_id is a ghost:
1390
check_parent = False
1391
# truncate the derived graph against this ghost.
1392
parent_text_key = None
1394
# look at the parent commit details inventories to
1395
# determine possible candidates in the per file graph.
1398
inv = inventory_cache[parent_id]
1400
inv = self.revision_tree(parent_id).inventory
1401
inventory_cache[parent_id] = inv
1402
parent_entry = inv._byid.get(text_key[0], None)
1403
if parent_entry is not None:
1405
text_key[0], parent_entry.revision)
1407
parent_text_key = None
1408
if parent_text_key is not None:
1409
candidate_parents.append(
1410
text_key_cache[parent_text_key])
1411
parent_heads = text_graph.heads(candidate_parents)
1412
new_parents = list(parent_heads)
1413
new_parents.sort(key=lambda x:candidate_parents.index(x))
1414
if new_parents == []:
1415
new_parents = [NULL_REVISION]
1416
text_index[text_key] = new_parents
1418
for text_key in invalid_keys:
1419
text_index[text_key] = [NULL_REVISION]
1422
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1423
"""Get an iterable listing the keys of all the data introduced by a set
1426
The keys will be ordered so that the corresponding items can be safely
1427
fetched and inserted in that order.
1429
:returns: An iterable producing tuples of (knit-kind, file-id,
1430
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1431
'revisions'. file-id is None unless knit-kind is 'file'.
1433
# XXX: it's a bit weird to control the inventory weave caching in this
1434
# generator. Ideally the caching would be done in fetch.py I think. Or
1435
# maybe this generator should explicitly have the contract that it
1436
# should not be iterated until the previously yielded item has been
1438
inv_w = self.inventories
1440
# file ids that changed
1441
file_ids = self.fileids_altered_by_revision_ids(revision_ids, inv_w)
1443
num_file_ids = len(file_ids)
1444
for file_id, altered_versions in file_ids.iteritems():
1445
if _files_pb is not None:
1446
_files_pb.update("fetch texts", count, num_file_ids)
1448
yield ("file", file_id, altered_versions)
1449
# We're done with the files_pb. Note that it finished by the caller,
1450
# just as it was created by the caller.
1454
yield ("inventory", None, revision_ids)
1457
revisions_with_signatures = set()
1458
for rev_id in revision_ids:
1460
self.get_signature_text(rev_id)
1461
except errors.NoSuchRevision:
1465
revisions_with_signatures.add(rev_id)
1466
yield ("signatures", None, revisions_with_signatures)
1469
yield ("revisions", None, revision_ids)
1472
def get_inventory(self, revision_id):
1473
"""Get Inventory object by revision id."""
1474
return self.iter_inventories([revision_id]).next()
1476
def iter_inventories(self, revision_ids):
1477
"""Get many inventories by revision_ids.
1479
This will buffer some or all of the texts used in constructing the
1480
inventories in memory, but will only parse a single inventory at a
1483
:return: An iterator of inventories.
1485
if ((None in revision_ids)
1486
or (_mod_revision.NULL_REVISION in revision_ids)):
1487
raise ValueError('cannot get null revision inventory')
1488
return self._iter_inventories(revision_ids)
1490
def _iter_inventories(self, revision_ids):
1491
"""single-document based inventory iteration."""
1492
for text, revision_id in self._iter_inventory_xmls(revision_ids):
1493
yield self.deserialise_inventory(revision_id, text)
1495
def _iter_inventory_xmls(self, revision_ids):
1496
keys = [(revision_id,) for revision_id in revision_ids]
1497
stream = self.inventories.get_record_stream(keys, 'unordered', True)
1499
for record in stream:
1500
if record.storage_kind != 'absent':
1501
texts[record.key] = record.get_bytes_as('fulltext')
1503
raise errors.NoSuchRevision(self, record.key)
1505
yield texts[key], key[-1]
1507
def deserialise_inventory(self, revision_id, xml):
1508
"""Transform the xml into an inventory object.
1510
:param revision_id: The expected revision id of the inventory.
1511
:param xml: A serialised inventory.
1513
result = self._serializer.read_inventory_from_string(xml, revision_id)
1514
if result.revision_id != revision_id:
1515
raise AssertionError('revision id mismatch %s != %s' % (
1516
result.revision_id, revision_id))
1519
def serialise_inventory(self, inv):
1520
return self._serializer.write_inventory_to_string(inv)
1522
def _serialise_inventory_to_lines(self, inv):
1523
return self._serializer.write_inventory_to_lines(inv)
1525
def get_serializer_format(self):
1526
return self._serializer.format_num
1529
def get_inventory_xml(self, revision_id):
1530
"""Get inventory XML as a file object."""
1531
texts = self._iter_inventory_xmls([revision_id])
1533
text, revision_id = texts.next()
1534
except StopIteration:
1535
raise errors.HistoryMissing(self, 'inventory', revision_id)
1539
def get_inventory_sha1(self, revision_id):
1540
"""Return the sha1 hash of the inventory entry
1542
return self.get_revision(revision_id).inventory_sha1
1544
def iter_reverse_revision_history(self, revision_id):
1545
"""Iterate backwards through revision ids in the lefthand history
1547
:param revision_id: The revision id to start with. All its lefthand
1548
ancestors will be traversed.
1550
graph = self.get_graph()
1551
next_id = revision_id
1553
if next_id in (None, _mod_revision.NULL_REVISION):
1556
# Note: The following line may raise KeyError in the event of
1557
# truncated history. We decided not to have a try:except:raise
1558
# RevisionNotPresent here until we see a use for it, because of the
1559
# cost in an inner loop that is by its very nature O(history).
1560
# Robert Collins 20080326
1561
parents = graph.get_parent_map([next_id])[next_id]
1562
if len(parents) == 0:
1565
next_id = parents[0]
1568
def get_revision_inventory(self, revision_id):
1569
"""Return inventory of a past revision."""
1570
# TODO: Unify this with get_inventory()
1571
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1572
# must be the same as its revision, so this is trivial.
1573
if revision_id is None:
1574
# This does not make sense: if there is no revision,
1575
# then it is the current tree inventory surely ?!
1576
# and thus get_root_id() is something that looks at the last
1577
# commit on the branch, and the get_root_id is an inventory check.
1578
raise NotImplementedError
1579
# return Inventory(self.get_root_id())
1581
return self.get_inventory(revision_id)
1584
def is_shared(self):
1585
"""Return True if this repository is flagged as a shared repository."""
1586
raise NotImplementedError(self.is_shared)
1589
def reconcile(self, other=None, thorough=False):
1590
"""Reconcile this repository."""
1591
from bzrlib.reconcile import RepoReconciler
1592
reconciler = RepoReconciler(self, thorough=thorough)
1593
reconciler.reconcile()
1596
def _refresh_data(self):
1597
"""Helper called from lock_* to ensure coherency with disk.
1599
The default implementation does nothing; it is however possible
1600
for repositories to maintain loaded indices across multiple locks
1601
by checking inside their implementation of this method to see
1602
whether their indices are still valid. This depends of course on
1603
the disk format being validatable in this manner.
1607
def revision_tree(self, revision_id):
1608
"""Return Tree for a revision on this branch.
1610
`revision_id` may be None for the empty tree revision.
1612
# TODO: refactor this to use an existing revision object
1613
# so we don't need to read it in twice.
1614
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1615
return RevisionTree(self, Inventory(root_id=None),
1616
_mod_revision.NULL_REVISION)
1618
inv = self.get_revision_inventory(revision_id)
1619
return RevisionTree(self, inv, revision_id)
1621
def revision_trees(self, revision_ids):
1622
"""Return Tree for a revision on this branch.
1624
`revision_id` may not be None or 'null:'"""
1625
inventories = self.iter_inventories(revision_ids)
1626
for inv in inventories:
1627
yield RevisionTree(self, inv, inv.revision_id)
1630
def get_ancestry(self, revision_id, topo_sorted=True):
1631
"""Return a list of revision-ids integrated by a revision.
1633
The first element of the list is always None, indicating the origin
1634
revision. This might change when we have history horizons, or
1635
perhaps we should have a new API.
1637
This is topologically sorted.
1639
if _mod_revision.is_null(revision_id):
1641
if not self.has_revision(revision_id):
1642
raise errors.NoSuchRevision(self, revision_id)
1643
graph = self.get_graph()
1645
search = graph._make_breadth_first_searcher([revision_id])
1648
found, ghosts = search.next_with_ghosts()
1649
except StopIteration:
1652
if _mod_revision.NULL_REVISION in keys:
1653
keys.remove(_mod_revision.NULL_REVISION)
1655
parent_map = graph.get_parent_map(keys)
1656
keys = tsort.topo_sort(parent_map)
1657
return [None] + list(keys)
1660
"""Compress the data within the repository.
1662
This operation only makes sense for some repository types. For other
1663
types it should be a no-op that just returns.
1665
This stub method does not require a lock, but subclasses should use
1666
@needs_write_lock as this is a long running call its reasonable to
1667
implicitly lock for the user.
1671
@deprecated_method(one_six)
1672
def print_file(self, file, revision_id):
1673
"""Print `file` to stdout.
1675
FIXME RBC 20060125 as John Meinel points out this is a bad api
1676
- it writes to stdout, it assumes that that is valid etc. Fix
1677
by creating a new more flexible convenience function.
1679
tree = self.revision_tree(revision_id)
1680
# use inventory as it was in that revision
1681
file_id = tree.inventory.path2id(file)
1683
# TODO: jam 20060427 Write a test for this code path
1684
# it had a bug in it, and was raising the wrong
1686
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1687
tree.print_file(file_id)
1689
def get_transaction(self):
1690
return self.control_files.get_transaction()
1692
@deprecated_method(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
if not self.has_revision(revision_id):
1773
raise errors.NoSuchRevision(self, revision_id)
1774
sig_present = (1 == len(
1775
self.signatures.get_parent_map([(revision_id,)])))
1779
def get_signature_text(self, revision_id):
1780
"""Return the text for a signature."""
1781
stream = self.signatures.get_record_stream([(revision_id,)],
1783
record = stream.next()
1784
if record.storage_kind == 'absent':
1785
raise errors.NoSuchRevision(self, revision_id)
1786
return record.get_bytes_as('fulltext')
1789
def check(self, revision_ids=None):
1790
"""Check consistency of all history of given revision_ids.
1792
Different repository implementations should override _check().
1794
:param revision_ids: A non-empty list of revision_ids whose ancestry
1795
will be checked. Typically the last revision_id of a branch.
1797
return self._check(revision_ids)
1799
def _check(self, revision_ids):
1800
result = check.Check(self)
1804
def _warn_if_deprecated(self):
1805
global _deprecation_warning_done
1806
if _deprecation_warning_done:
1808
_deprecation_warning_done = True
1809
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1810
% (self._format, self.bzrdir.transport.base))
1812
def supports_rich_root(self):
1813
return self._format.rich_root_data
1815
def _check_ascii_revisionid(self, revision_id, method):
1816
"""Private helper for ascii-only repositories."""
1817
# weave repositories refuse to store revisionids that are non-ascii.
1818
if revision_id is not None:
1819
# weaves require ascii revision ids.
1820
if isinstance(revision_id, unicode):
1822
revision_id.encode('ascii')
1823
except UnicodeEncodeError:
1824
raise errors.NonAsciiRevisionId(method, self)
1827
revision_id.decode('ascii')
1828
except UnicodeDecodeError:
1829
raise errors.NonAsciiRevisionId(method, self)
1831
def revision_graph_can_have_wrong_parents(self):
1832
"""Is it possible for this repository to have a revision graph with
1835
If True, then this repository must also implement
1836
_find_inconsistent_revision_parents so that check and reconcile can
1837
check for inconsistencies before proceeding with other checks that may
1838
depend on the revision index being consistent.
1840
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1843
# remove these delegates a while after bzr 0.15
1844
def __make_delegated(name, from_module):
1845
def _deprecated_repository_forwarder():
1846
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1847
% (name, from_module),
1850
m = __import__(from_module, globals(), locals(), [name])
1852
return getattr(m, name)
1853
except AttributeError:
1854
raise AttributeError('module %s has no name %s'
1856
globals()[name] = _deprecated_repository_forwarder
1859
'AllInOneRepository',
1860
'WeaveMetaDirRepository',
1861
'PreSplitOutRepositoryFormat',
1862
'RepositoryFormat4',
1863
'RepositoryFormat5',
1864
'RepositoryFormat6',
1865
'RepositoryFormat7',
1867
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1871
'RepositoryFormatKnit',
1872
'RepositoryFormatKnit1',
1874
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1877
def install_revision(repository, rev, revision_tree):
1878
"""Install all revision data into a repository."""
1879
install_revisions(repository, [(rev, revision_tree, None)])
1882
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1883
"""Install all revision data into a repository.
1885
Accepts an iterable of revision, tree, signature tuples. The signature
1888
repository.start_write_group()
1890
for n, (revision, revision_tree, signature) in enumerate(iterable):
1891
_install_revision(repository, revision, revision_tree, signature)
1893
pb.update('Transferring revisions', n + 1, num_revisions)
1895
repository.abort_write_group()
1898
repository.commit_write_group()
1901
def _install_revision(repository, rev, revision_tree, signature):
1902
"""Install all revision data into a repository."""
1903
present_parents = []
1905
for p_id in rev.parent_ids:
1906
if repository.has_revision(p_id):
1907
present_parents.append(p_id)
1908
parent_trees[p_id] = repository.revision_tree(p_id)
1910
parent_trees[p_id] = repository.revision_tree(None)
1912
inv = revision_tree.inventory
1913
entries = inv.iter_entries()
1914
# backwards compatibility hack: skip the root id.
1915
if not repository.supports_rich_root():
1916
path, root = entries.next()
1917
if root.revision != rev.revision_id:
1918
raise errors.IncompatibleRevision(repr(repository))
1920
for path, ie in entries:
1921
text_keys[(ie.file_id, ie.revision)] = ie
1922
text_parent_map = repository.texts.get_parent_map(text_keys)
1923
missing_texts = set(text_keys) - set(text_parent_map)
1924
# Add the texts that are not already present
1925
for text_key in missing_texts:
1926
ie = text_keys[text_key]
1928
# FIXME: TODO: The following loop overlaps/duplicates that done by
1929
# commit to determine parents. There is a latent/real bug here where
1930
# the parents inserted are not those commit would do - in particular
1931
# they are not filtered by heads(). RBC, AB
1932
for revision, tree in parent_trees.iteritems():
1933
if ie.file_id not in tree:
1935
parent_id = tree.inventory[ie.file_id].revision
1936
if parent_id in text_parents:
1938
text_parents.append((ie.file_id, parent_id))
1939
lines = revision_tree.get_file(ie.file_id).readlines()
1940
repository.texts.add_lines(text_key, text_parents, lines)
1942
# install the inventory
1943
repository.add_inventory(rev.revision_id, inv, present_parents)
1944
except errors.RevisionAlreadyPresent:
1946
if signature is not None:
1947
repository.add_signature_text(rev.revision_id, signature)
1948
repository.add_revision(rev.revision_id, rev, inv)
1951
class MetaDirRepository(Repository):
1952
"""Repositories in the new meta-dir layout.
1954
:ivar _transport: Transport for access to repository control files,
1955
typically pointing to .bzr/repository.
1958
def __init__(self, _format, a_bzrdir, control_files):
1959
super(MetaDirRepository, self).__init__(_format, a_bzrdir, control_files)
1960
self._transport = control_files._transport
1963
def is_shared(self):
1964
"""Return True if this repository is flagged as a shared repository."""
1965
return self._transport.has('shared-storage')
1968
def set_make_working_trees(self, new_value):
1969
"""Set the policy flag for making working trees when creating branches.
1971
This only applies to branches that use this repository.
1973
The default is 'True'.
1974
:param new_value: True to restore the default, False to disable making
1979
self._transport.delete('no-working-trees')
1980
except errors.NoSuchFile:
1983
self._transport.put_bytes('no-working-trees', '',
1984
mode=self.bzrdir._get_file_mode())
1986
def make_working_trees(self):
1987
"""Returns the policy for making working trees on new branches."""
1988
return not self._transport.has('no-working-trees')
1991
class MetaDirVersionedFileRepository(MetaDirRepository):
1992
"""Repositories in a meta-dir, that work via versioned file objects."""
1994
def __init__(self, _format, a_bzrdir, control_files):
1995
super(MetaDirVersionedFileRepository, self).__init__(_format, a_bzrdir,
1999
class RepositoryFormatRegistry(registry.Registry):
2000
"""Registry of RepositoryFormats."""
2002
def get(self, format_string):
2003
r = registry.Registry.get(self, format_string)
2009
format_registry = RepositoryFormatRegistry()
2010
"""Registry of formats, indexed by their identifying format string.
2012
This can contain either format instances themselves, or classes/factories that
2013
can be called to obtain one.
2017
#####################################################################
2018
# Repository Formats
2020
class RepositoryFormat(object):
2021
"""A repository format.
2023
Formats provide three things:
2024
* An initialization routine to construct repository data on disk.
2025
* a format string which is used when the BzrDir supports versioned
2027
* an open routine which returns a Repository instance.
2029
There is one and only one Format subclass for each on-disk format. But
2030
there can be one Repository subclass that is used for several different
2031
formats. The _format attribute on a Repository instance can be used to
2032
determine the disk format.
2034
Formats are placed in an dict by their format string for reference
2035
during opening. These should be subclasses of RepositoryFormat
2038
Once a format is deprecated, just deprecate the initialize and open
2039
methods on the format class. Do not deprecate the object, as the
2040
object will be created every system load.
2042
Common instance attributes:
2043
_matchingbzrdir - the bzrdir format that the repository format was
2044
originally written to work with. This can be used if manually
2045
constructing a bzrdir and repository, or more commonly for test suite
2049
# Set to True or False in derived classes. True indicates that the format
2050
# supports ghosts gracefully.
2051
supports_ghosts = None
2052
# Can this repository be given external locations to lookup additional
2053
# data. Set to True or False in derived classes.
2054
supports_external_lookups = None
2057
return "<%s>" % self.__class__.__name__
2059
def __eq__(self, other):
2060
# format objects are generally stateless
2061
return isinstance(other, self.__class__)
2063
def __ne__(self, other):
2064
return not self == other
2067
def find_format(klass, a_bzrdir):
2068
"""Return the format for the repository object in a_bzrdir.
2070
This is used by bzr native formats that have a "format" file in
2071
the repository. Other methods may be used by different types of
2075
transport = a_bzrdir.get_repository_transport(None)
2076
format_string = transport.get("format").read()
2077
return format_registry.get(format_string)
2078
except errors.NoSuchFile:
2079
raise errors.NoRepositoryPresent(a_bzrdir)
2081
raise errors.UnknownFormatError(format=format_string,
2085
def register_format(klass, format):
2086
format_registry.register(format.get_format_string(), format)
2089
def unregister_format(klass, format):
2090
format_registry.remove(format.get_format_string())
2093
def get_default_format(klass):
2094
"""Return the current default format."""
2095
from bzrlib import bzrdir
2096
return bzrdir.format_registry.make_bzrdir('default').repository_format
2098
def get_format_string(self):
2099
"""Return the ASCII format string that identifies this format.
2101
Note that in pre format ?? repositories the format string is
2102
not permitted nor written to disk.
2104
raise NotImplementedError(self.get_format_string)
2106
def get_format_description(self):
2107
"""Return the short description for this format."""
2108
raise NotImplementedError(self.get_format_description)
2110
# TODO: this shouldn't be in the base class, it's specific to things that
2111
# use weaves or knits -- mbp 20070207
2112
def _get_versioned_file_store(self,
2117
versionedfile_class=None,
2118
versionedfile_kwargs={},
2120
if versionedfile_class is None:
2121
versionedfile_class = self._versionedfile_class
2122
weave_transport = control_files._transport.clone(name)
2123
dir_mode = control_files._dir_mode
2124
file_mode = control_files._file_mode
2125
return VersionedFileStore(weave_transport, prefixed=prefixed,
2127
file_mode=file_mode,
2128
versionedfile_class=versionedfile_class,
2129
versionedfile_kwargs=versionedfile_kwargs,
2132
def initialize(self, a_bzrdir, shared=False):
2133
"""Initialize a repository of this format in a_bzrdir.
2135
:param a_bzrdir: The bzrdir to put the new repository in it.
2136
:param shared: The repository should be initialized as a sharable one.
2137
:returns: The new repository object.
2139
This may raise UninitializableFormat if shared repository are not
2140
compatible the a_bzrdir.
2142
raise NotImplementedError(self.initialize)
2144
def is_supported(self):
2145
"""Is this format supported?
2147
Supported formats must be initializable and openable.
2148
Unsupported formats may not support initialization or committing or
2149
some other features depending on the reason for not being supported.
2153
def check_conversion_target(self, target_format):
2154
raise NotImplementedError(self.check_conversion_target)
2156
def open(self, a_bzrdir, _found=False):
2157
"""Return an instance of this format for the bzrdir a_bzrdir.
2159
_found is a private parameter, do not use it.
2161
raise NotImplementedError(self.open)
2164
class MetaDirRepositoryFormat(RepositoryFormat):
2165
"""Common base class for the new repositories using the metadir layout."""
2167
rich_root_data = False
2168
supports_tree_reference = False
2169
supports_external_lookups = False
2170
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
2173
super(MetaDirRepositoryFormat, self).__init__()
2175
def _create_control_files(self, a_bzrdir):
2176
"""Create the required files and the initial control_files object."""
2177
# FIXME: RBC 20060125 don't peek under the covers
2178
# NB: no need to escape relative paths that are url safe.
2179
repository_transport = a_bzrdir.get_repository_transport(self)
2180
control_files = lockable_files.LockableFiles(repository_transport,
2181
'lock', lockdir.LockDir)
2182
control_files.create_lock()
2183
return control_files
2185
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
2186
"""Upload the initial blank content."""
2187
control_files = self._create_control_files(a_bzrdir)
2188
control_files.lock_write()
2189
transport = control_files._transport
2191
utf8_files += [('shared-storage', '')]
2193
transport.mkdir_multi(dirs, mode=a_bzrdir._get_dir_mode())
2194
for (filename, content_stream) in files:
2195
transport.put_file(filename, content_stream,
2196
mode=a_bzrdir._get_file_mode())
2197
for (filename, content_bytes) in utf8_files:
2198
transport.put_bytes_non_atomic(filename, content_bytes,
2199
mode=a_bzrdir._get_file_mode())
2201
control_files.unlock()
2204
# formats which have no format string are not discoverable
2205
# and not independently creatable, so are not registered. They're
2206
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
2207
# needed, it's constructed directly by the BzrDir. Non-native formats where
2208
# the repository is not separately opened are similar.
2210
format_registry.register_lazy(
2211
'Bazaar-NG Repository format 7',
2212
'bzrlib.repofmt.weaverepo',
2216
format_registry.register_lazy(
2217
'Bazaar-NG Knit Repository Format 1',
2218
'bzrlib.repofmt.knitrepo',
2219
'RepositoryFormatKnit1',
2222
format_registry.register_lazy(
2223
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
2224
'bzrlib.repofmt.knitrepo',
2225
'RepositoryFormatKnit3',
2228
format_registry.register_lazy(
2229
'Bazaar Knit Repository Format 4 (bzr 1.0)\n',
2230
'bzrlib.repofmt.knitrepo',
2231
'RepositoryFormatKnit4',
2234
# Pack-based formats. There is one format for pre-subtrees, and one for
2235
# post-subtrees to allow ease of testing.
2236
# NOTE: These are experimental in 0.92. Stable in 1.0 and above
2237
format_registry.register_lazy(
2238
'Bazaar pack repository format 1 (needs bzr 0.92)\n',
2239
'bzrlib.repofmt.pack_repo',
2240
'RepositoryFormatKnitPack1',
2242
format_registry.register_lazy(
2243
'Bazaar pack repository format 1 with subtree support (needs bzr 0.92)\n',
2244
'bzrlib.repofmt.pack_repo',
2245
'RepositoryFormatKnitPack3',
2247
format_registry.register_lazy(
2248
'Bazaar pack repository format 1 with rich root (needs bzr 1.0)\n',
2249
'bzrlib.repofmt.pack_repo',
2250
'RepositoryFormatKnitPack4',
2252
# Development formats.
2254
# development 0 - stub to introduce development versioning scheme.
2255
format_registry.register_lazy(
2256
"Bazaar development format 0 (needs bzr.dev from before 1.3)\n",
2257
'bzrlib.repofmt.pack_repo',
2258
'RepositoryFormatPackDevelopment0',
2260
format_registry.register_lazy(
2261
("Bazaar development format 0 with subtree support "
2262
"(needs bzr.dev from before 1.3)\n"),
2263
'bzrlib.repofmt.pack_repo',
2264
'RepositoryFormatPackDevelopment0Subtree',
2266
# 1.3->1.4 go below here
2269
class InterRepository(InterObject):
2270
"""This class represents operations taking place between two repositories.
2272
Its instances have methods like copy_content and fetch, and contain
2273
references to the source and target repositories these operations can be
2276
Often we will provide convenience methods on 'repository' which carry out
2277
operations with another repository - they will always forward to
2278
InterRepository.get(other).method_name(parameters).
2282
"""The available optimised InterRepository types."""
2284
def copy_content(self, revision_id=None):
2285
raise NotImplementedError(self.copy_content)
2287
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2288
"""Fetch the content required to construct revision_id.
2290
The content is copied from self.source to self.target.
2292
:param revision_id: if None all content is copied, if NULL_REVISION no
2294
:param pb: optional progress bar to use for progress reports. If not
2295
provided a default one will be created.
2297
Returns the copied revision count and the failed revisions in a tuple:
2300
raise NotImplementedError(self.fetch)
2302
def _walk_to_common_revisions(self, revision_ids):
2303
"""Walk out from revision_ids in source to revisions target has.
2305
:param revision_ids: The start point for the search.
2306
:return: A set of revision ids.
2308
target_graph = self.target.get_graph()
2309
revision_ids = frozenset(revision_ids)
2310
if set(target_graph.get_parent_map(revision_ids)) == revision_ids:
2311
return graph.SearchResult(revision_ids, set(), 0, set())
2312
missing_revs = set()
2313
source_graph = self.source.get_graph()
2314
# ensure we don't pay silly lookup costs.
2315
searcher = source_graph._make_breadth_first_searcher(revision_ids)
2316
null_set = frozenset([_mod_revision.NULL_REVISION])
2319
next_revs, ghosts = searcher.next_with_ghosts()
2320
except StopIteration:
2322
if revision_ids.intersection(ghosts):
2323
absent_ids = set(revision_ids.intersection(ghosts))
2324
# If all absent_ids are present in target, no error is needed.
2325
absent_ids.difference_update(
2326
set(target_graph.get_parent_map(absent_ids)))
2328
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2329
# we don't care about other ghosts as we can't fetch them and
2330
# haven't been asked to.
2331
next_revs = set(next_revs)
2332
# we always have NULL_REVISION present.
2333
have_revs = set(target_graph.get_parent_map(next_revs)).union(null_set)
2334
missing_revs.update(next_revs - have_revs)
2335
searcher.stop_searching_any(have_revs)
2336
return searcher.get_result()
2338
@deprecated_method(one_two)
2340
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2341
"""Return the revision ids that source has that target does not.
2343
These are returned in topological order.
2345
:param revision_id: only return revision ids included by this
2347
:param find_ghosts: If True find missing revisions in deep history
2348
rather than just finding the surface difference.
2350
return list(self.search_missing_revision_ids(
2351
revision_id, find_ghosts).get_keys())
2354
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2355
"""Return the revision ids that source has that target does not.
2357
:param revision_id: only return revision ids included by this
2359
:param find_ghosts: If True find missing revisions in deep history
2360
rather than just finding the surface difference.
2361
:return: A bzrlib.graph.SearchResult.
2363
# stop searching at found target revisions.
2364
if not find_ghosts and revision_id is not None:
2365
return self._walk_to_common_revisions([revision_id])
2366
# generic, possibly worst case, slow code path.
2367
target_ids = set(self.target.all_revision_ids())
2368
if revision_id is not None:
2369
source_ids = self.source.get_ancestry(revision_id)
2370
if source_ids[0] is not None:
2371
raise AssertionError()
2374
source_ids = self.source.all_revision_ids()
2375
result_set = set(source_ids).difference(target_ids)
2376
return self.source.revision_ids_to_search_result(result_set)
2379
def _same_model(source, target):
2380
"""True if source and target have the same data representation."""
2381
if source.supports_rich_root() != target.supports_rich_root():
2383
if source._serializer != target._serializer:
2388
class InterSameDataRepository(InterRepository):
2389
"""Code for converting between repositories that represent the same data.
2391
Data format and model must match for this to work.
2395
def _get_repo_format_to_test(self):
2396
"""Repository format for testing with.
2398
InterSameData can pull from subtree to subtree and from non-subtree to
2399
non-subtree, so we test this with the richest repository format.
2401
from bzrlib.repofmt import knitrepo
2402
return knitrepo.RepositoryFormatKnit3()
2405
def is_compatible(source, target):
2406
return InterRepository._same_model(source, target)
2409
def copy_content(self, revision_id=None):
2410
"""Make a complete copy of the content in self into destination.
2412
This copies both the repository's revision data, and configuration information
2413
such as the make_working_trees setting.
2415
This is a destructive operation! Do not use it on existing
2418
:param revision_id: Only copy the content needed to construct
2419
revision_id and its parents.
2422
self.target.set_make_working_trees(self.source.make_working_trees())
2423
except NotImplementedError:
2425
# but don't bother fetching if we have the needed data now.
2426
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2427
self.target.has_revision(revision_id)):
2429
self.target.fetch(self.source, revision_id=revision_id)
2432
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2433
"""See InterRepository.fetch()."""
2434
from bzrlib.fetch import GenericRepoFetcher
2435
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2436
self.source, self.source._format, self.target,
2437
self.target._format)
2438
f = GenericRepoFetcher(to_repository=self.target,
2439
from_repository=self.source,
2440
last_revision=revision_id,
2441
pb=pb, find_ghosts=find_ghosts)
2442
return f.count_copied, f.failed_revisions
2445
class InterWeaveRepo(InterSameDataRepository):
2446
"""Optimised code paths between Weave based repositories.
2448
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2449
implemented lazy inter-object optimisation.
2453
def _get_repo_format_to_test(self):
2454
from bzrlib.repofmt import weaverepo
2455
return weaverepo.RepositoryFormat7()
2458
def is_compatible(source, target):
2459
"""Be compatible with known Weave formats.
2461
We don't test for the stores being of specific types because that
2462
could lead to confusing results, and there is no need to be
2465
from bzrlib.repofmt.weaverepo import (
2471
return (isinstance(source._format, (RepositoryFormat5,
2473
RepositoryFormat7)) and
2474
isinstance(target._format, (RepositoryFormat5,
2476
RepositoryFormat7)))
2477
except AttributeError:
2481
def copy_content(self, revision_id=None):
2482
"""See InterRepository.copy_content()."""
2483
# weave specific optimised path:
2485
self.target.set_make_working_trees(self.source.make_working_trees())
2486
except (errors.RepositoryUpgradeRequired, NotImplemented):
2488
# FIXME do not peek!
2489
if self.source._transport.listable():
2490
pb = ui.ui_factory.nested_progress_bar()
2492
self.target.texts.insert_record_stream(
2493
self.source.texts.get_record_stream(
2494
self.source.texts.keys(), 'topological', False))
2495
pb.update('copying inventory', 0, 1)
2496
self.target.inventories.insert_record_stream(
2497
self.source.inventories.get_record_stream(
2498
self.source.inventories.keys(), 'topological', False))
2499
self.target.signatures.insert_record_stream(
2500
self.source.signatures.get_record_stream(
2501
self.source.signatures.keys(),
2503
self.target.revisions.insert_record_stream(
2504
self.source.revisions.get_record_stream(
2505
self.source.revisions.keys(),
2506
'topological', True))
2510
self.target.fetch(self.source, revision_id=revision_id)
2513
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2514
"""See InterRepository.fetch()."""
2515
from bzrlib.fetch import GenericRepoFetcher
2516
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2517
self.source, self.source._format, self.target, self.target._format)
2518
f = GenericRepoFetcher(to_repository=self.target,
2519
from_repository=self.source,
2520
last_revision=revision_id,
2521
pb=pb, find_ghosts=find_ghosts)
2522
return f.count_copied, f.failed_revisions
2525
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2526
"""See InterRepository.missing_revision_ids()."""
2527
# we want all revisions to satisfy revision_id in source.
2528
# but we don't want to stat every file here and there.
2529
# we want then, all revisions other needs to satisfy revision_id
2530
# checked, but not those that we have locally.
2531
# so the first thing is to get a subset of the revisions to
2532
# satisfy revision_id in source, and then eliminate those that
2533
# we do already have.
2534
# this is slow on high latency connection to self, but as as this
2535
# disk format scales terribly for push anyway due to rewriting
2536
# inventory.weave, this is considered acceptable.
2538
if revision_id is not None:
2539
source_ids = self.source.get_ancestry(revision_id)
2540
if source_ids[0] is not None:
2541
raise AssertionError()
2544
source_ids = self.source._all_possible_ids()
2545
source_ids_set = set(source_ids)
2546
# source_ids is the worst possible case we may need to pull.
2547
# now we want to filter source_ids against what we actually
2548
# have in target, but don't try to check for existence where we know
2549
# we do not have a revision as that would be pointless.
2550
target_ids = set(self.target._all_possible_ids())
2551
possibly_present_revisions = target_ids.intersection(source_ids_set)
2552
actually_present_revisions = set(
2553
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2554
required_revisions = source_ids_set.difference(actually_present_revisions)
2555
if revision_id is not None:
2556
# we used get_ancestry to determine source_ids then we are assured all
2557
# revisions referenced are present as they are installed in topological order.
2558
# and the tip revision was validated by get_ancestry.
2559
result_set = required_revisions
2561
# if we just grabbed the possibly available ids, then
2562
# we only have an estimate of whats available and need to validate
2563
# that against the revision records.
2565
self.source._eliminate_revisions_not_present(required_revisions))
2566
return self.source.revision_ids_to_search_result(result_set)
2569
class InterKnitRepo(InterSameDataRepository):
2570
"""Optimised code paths between Knit based repositories."""
2573
def _get_repo_format_to_test(self):
2574
from bzrlib.repofmt import knitrepo
2575
return knitrepo.RepositoryFormatKnit1()
2578
def is_compatible(source, target):
2579
"""Be compatible with known Knit formats.
2581
We don't test for the stores being of specific types because that
2582
could lead to confusing results, and there is no need to be
2585
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2587
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2588
isinstance(target._format, RepositoryFormatKnit))
2589
except AttributeError:
2591
return are_knits and InterRepository._same_model(source, target)
2594
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2595
"""See InterRepository.fetch()."""
2596
from bzrlib.fetch import KnitRepoFetcher
2597
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2598
self.source, self.source._format, self.target, self.target._format)
2599
f = KnitRepoFetcher(to_repository=self.target,
2600
from_repository=self.source,
2601
last_revision=revision_id,
2602
pb=pb, find_ghosts=find_ghosts)
2603
return f.count_copied, f.failed_revisions
2606
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2607
"""See InterRepository.missing_revision_ids()."""
2608
if revision_id is not None:
2609
source_ids = self.source.get_ancestry(revision_id)
2610
if source_ids[0] is not None:
2611
raise AssertionError()
2614
source_ids = self.source.all_revision_ids()
2615
source_ids_set = set(source_ids)
2616
# source_ids is the worst possible case we may need to pull.
2617
# now we want to filter source_ids against what we actually
2618
# have in target, but don't try to check for existence where we know
2619
# we do not have a revision as that would be pointless.
2620
target_ids = set(self.target.all_revision_ids())
2621
possibly_present_revisions = target_ids.intersection(source_ids_set)
2622
actually_present_revisions = set(
2623
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2624
required_revisions = source_ids_set.difference(actually_present_revisions)
2625
if revision_id is not None:
2626
# we used get_ancestry to determine source_ids then we are assured all
2627
# revisions referenced are present as they are installed in topological order.
2628
# and the tip revision was validated by get_ancestry.
2629
result_set = required_revisions
2631
# if we just grabbed the possibly available ids, then
2632
# we only have an estimate of whats available and need to validate
2633
# that against the revision records.
2635
self.source._eliminate_revisions_not_present(required_revisions))
2636
return self.source.revision_ids_to_search_result(result_set)
2639
class InterPackRepo(InterSameDataRepository):
2640
"""Optimised code paths between Pack based repositories."""
2643
def _get_repo_format_to_test(self):
2644
from bzrlib.repofmt import pack_repo
2645
return pack_repo.RepositoryFormatKnitPack1()
2648
def is_compatible(source, target):
2649
"""Be compatible with known Pack formats.
2651
We don't test for the stores being of specific types because that
2652
could lead to confusing results, and there is no need to be
2655
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2657
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2658
isinstance(target._format, RepositoryFormatPack))
2659
except AttributeError:
2661
return are_packs and InterRepository._same_model(source, target)
2664
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2665
"""See InterRepository.fetch()."""
2666
from bzrlib.repofmt.pack_repo import Packer
2667
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2668
self.source, self.source._format, self.target, self.target._format)
2669
self.count_copied = 0
2670
if revision_id is None:
2672
# everything to do - use pack logic
2673
# to fetch from all packs to one without
2674
# inventory parsing etc, IFF nothing to be copied is in the target.
2676
revision_ids = self.source.all_revision_ids()
2677
revision_keys = [(revid,) for revid in revision_ids]
2678
index = self.target._pack_collection.revision_index.combined_index
2679
present_revision_ids = set(item[1][0] for item in
2680
index.iter_entries(revision_keys))
2681
revision_ids = set(revision_ids) - present_revision_ids
2682
# implementing the TODO will involve:
2683
# - detecting when all of a pack is selected
2684
# - avoiding as much as possible pre-selection, so the
2685
# more-core routines such as create_pack_from_packs can filter in
2686
# a just-in-time fashion. (though having a HEADS list on a
2687
# repository might make this a lot easier, because we could
2688
# sensibly detect 'new revisions' without doing a full index scan.
2689
elif _mod_revision.is_null(revision_id):
2694
revision_ids = self.search_missing_revision_ids(revision_id,
2695
find_ghosts=find_ghosts).get_keys()
2696
except errors.NoSuchRevision:
2697
raise errors.InstallFailed([revision_id])
2698
if len(revision_ids) == 0:
2700
packs = self.source._pack_collection.all_packs()
2701
pack = Packer(self.target._pack_collection, packs, '.fetch',
2702
revision_ids).pack()
2703
if pack is not None:
2704
self.target._pack_collection._save_pack_names()
2705
# Trigger an autopack. This may duplicate effort as we've just done
2706
# a pack creation, but for now it is simpler to think about as
2707
# 'upload data, then repack if needed'.
2708
self.target._pack_collection.autopack()
2709
return (pack.get_revision_count(), [])
2714
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2715
"""See InterRepository.missing_revision_ids().
2717
:param find_ghosts: Find ghosts throughout the ancestry of
2720
if not find_ghosts and revision_id is not None:
2721
return self._walk_to_common_revisions([revision_id])
2722
elif revision_id is not None:
2723
source_ids = self.source.get_ancestry(revision_id)
2724
if source_ids[0] is not None:
2725
raise AssertionError()
2728
source_ids = self.source.all_revision_ids()
2729
# source_ids is the worst possible case we may need to pull.
2730
# now we want to filter source_ids against what we actually
2731
# have in target, but don't try to check for existence where we know
2732
# we do not have a revision as that would be pointless.
2733
target_ids = set(self.target.all_revision_ids())
2734
result_set = set(source_ids).difference(target_ids)
2735
return self.source.revision_ids_to_search_result(result_set)
2738
class InterModel1and2(InterRepository):
2741
def _get_repo_format_to_test(self):
2745
def is_compatible(source, target):
2746
if not source.supports_rich_root() and target.supports_rich_root():
2752
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2753
"""See InterRepository.fetch()."""
2754
from bzrlib.fetch import Model1toKnit2Fetcher
2755
f = Model1toKnit2Fetcher(to_repository=self.target,
2756
from_repository=self.source,
2757
last_revision=revision_id,
2758
pb=pb, find_ghosts=find_ghosts)
2759
return f.count_copied, f.failed_revisions
2762
def copy_content(self, revision_id=None):
2763
"""Make a complete copy of the content in self into destination.
2765
This is a destructive operation! Do not use it on existing
2768
:param revision_id: Only copy the content needed to construct
2769
revision_id and its parents.
2772
self.target.set_make_working_trees(self.source.make_working_trees())
2773
except NotImplementedError:
2775
# but don't bother fetching if we have the needed data now.
2776
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2777
self.target.has_revision(revision_id)):
2779
self.target.fetch(self.source, revision_id=revision_id)
2782
class InterKnit1and2(InterKnitRepo):
2785
def _get_repo_format_to_test(self):
2789
def is_compatible(source, target):
2790
"""Be compatible with Knit1 source and Knit3 target"""
2791
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2793
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2794
RepositoryFormatKnit3)
2795
from bzrlib.repofmt.pack_repo import (
2796
RepositoryFormatKnitPack1,
2797
RepositoryFormatKnitPack3,
2798
RepositoryFormatPackDevelopment0,
2799
RepositoryFormatPackDevelopment0Subtree,
2802
RepositoryFormatKnit1,
2803
RepositoryFormatKnitPack1,
2804
RepositoryFormatPackDevelopment0,
2807
RepositoryFormatKnit3,
2808
RepositoryFormatKnitPack3,
2809
RepositoryFormatPackDevelopment0Subtree,
2811
return (isinstance(source._format, nosubtrees) and
2812
isinstance(target._format, subtrees))
2813
except AttributeError:
2817
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2818
"""See InterRepository.fetch()."""
2819
from bzrlib.fetch import Knit1to2Fetcher
2820
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2821
self.source, self.source._format, self.target,
2822
self.target._format)
2823
f = Knit1to2Fetcher(to_repository=self.target,
2824
from_repository=self.source,
2825
last_revision=revision_id,
2826
pb=pb, find_ghosts=find_ghosts)
2827
return f.count_copied, f.failed_revisions
2830
class InterDifferingSerializer(InterKnitRepo):
2833
def _get_repo_format_to_test(self):
2837
def is_compatible(source, target):
2838
"""Be compatible with Knit2 source and Knit3 target"""
2839
if source.supports_rich_root() != target.supports_rich_root():
2841
# Ideally, we'd support fetching if the source had no tree references
2842
# even if it supported them...
2843
if (getattr(source, '_format.supports_tree_reference', False) and
2844
not getattr(target, '_format.supports_tree_reference', False)):
2849
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2850
"""See InterRepository.fetch()."""
2851
revision_ids = self.target.search_missing_revision_ids(self.source,
2852
revision_id, find_ghosts=find_ghosts).get_keys()
2853
revision_ids = tsort.topo_sort(
2854
self.source.get_graph().get_parent_map(revision_ids))
2855
def revisions_iterator():
2856
for current_revision_id in revision_ids:
2857
revision = self.source.get_revision(current_revision_id)
2858
tree = self.source.revision_tree(current_revision_id)
2860
signature = self.source.get_signature_text(
2861
current_revision_id)
2862
except errors.NoSuchRevision:
2864
yield revision, tree, signature
2866
my_pb = ui.ui_factory.nested_progress_bar()
2871
install_revisions(self.target, revisions_iterator(),
2872
len(revision_ids), pb)
2874
if my_pb is not None:
2876
return len(revision_ids), 0
2879
class InterOtherToRemote(InterRepository):
2881
def __init__(self, source, target):
2882
InterRepository.__init__(self, source, target)
2883
self._real_inter = None
2886
def is_compatible(source, target):
2887
if isinstance(target, remote.RemoteRepository):
2891
def _ensure_real_inter(self):
2892
if self._real_inter is None:
2893
self.target._ensure_real()
2894
real_target = self.target._real_repository
2895
self._real_inter = InterRepository.get(self.source, real_target)
2897
def copy_content(self, revision_id=None):
2898
self._ensure_real_inter()
2899
self._real_inter.copy_content(revision_id=revision_id)
2901
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2902
self._ensure_real_inter()
2903
return self._real_inter.fetch(revision_id=revision_id, pb=pb,
2904
find_ghosts=find_ghosts)
2907
def _get_repo_format_to_test(self):
2911
InterRepository.register_optimiser(InterDifferingSerializer)
2912
InterRepository.register_optimiser(InterSameDataRepository)
2913
InterRepository.register_optimiser(InterWeaveRepo)
2914
InterRepository.register_optimiser(InterKnitRepo)
2915
InterRepository.register_optimiser(InterModel1and2)
2916
InterRepository.register_optimiser(InterKnit1and2)
2917
InterRepository.register_optimiser(InterPackRepo)
2918
InterRepository.register_optimiser(InterOtherToRemote)
2921
class CopyConverter(object):
2922
"""A repository conversion tool which just performs a copy of the content.
2924
This is slow but quite reliable.
2927
def __init__(self, target_format):
2928
"""Create a CopyConverter.
2930
:param target_format: The format the resulting repository should be.
2932
self.target_format = target_format
2934
def convert(self, repo, pb):
2935
"""Perform the conversion of to_convert, giving feedback via pb.
2937
:param to_convert: The disk object to convert.
2938
:param pb: a progress bar to use for progress information.
2943
# this is only useful with metadir layouts - separated repo content.
2944
# trigger an assertion if not such
2945
repo._format.get_format_string()
2946
self.repo_dir = repo.bzrdir
2947
self.step('Moving repository to repository.backup')
2948
self.repo_dir.transport.move('repository', 'repository.backup')
2949
backup_transport = self.repo_dir.transport.clone('repository.backup')
2950
repo._format.check_conversion_target(self.target_format)
2951
self.source_repo = repo._format.open(self.repo_dir,
2953
_override_transport=backup_transport)
2954
self.step('Creating new repository')
2955
converted = self.target_format.initialize(self.repo_dir,
2956
self.source_repo.is_shared())
2957
converted.lock_write()
2959
self.step('Copying content into repository.')
2960
self.source_repo.copy_content_into(converted)
2963
self.step('Deleting old repository content.')
2964
self.repo_dir.transport.delete_tree('repository.backup')
2965
self.pb.note('repository converted')
2967
def step(self, message):
2968
"""Update the pb by a step."""
2970
self.pb.update(message, self.count, self.total)
2982
def _unescaper(match, _map=_unescape_map):
2983
code = match.group(1)
2987
if not code.startswith('#'):
2989
return unichr(int(code[1:])).encode('utf8')
2995
def _unescape_xml(data):
2996
"""Unescape predefined XML entities in a string of data."""
2998
if _unescape_re is None:
2999
_unescape_re = re.compile('\&([^;]*);')
3000
return _unescape_re.sub(_unescaper, data)
3003
class _VersionedFileChecker(object):
3005
def __init__(self, repository):
3006
self.repository = repository
3007
self.text_index = self.repository._generate_text_key_index()
3009
def calculate_file_version_parents(self, text_key):
3010
"""Calculate the correct parents for a file version according to
3013
parent_keys = self.text_index[text_key]
3014
if parent_keys == [_mod_revision.NULL_REVISION]:
3016
return tuple(parent_keys)
3018
def check_file_version_parents(self, texts, progress_bar=None):
3019
"""Check the parents stored in a versioned file are correct.
3021
It also detects file versions that are not referenced by their
3022
corresponding revision's inventory.
3024
:returns: A tuple of (wrong_parents, dangling_file_versions).
3025
wrong_parents is a dict mapping {revision_id: (stored_parents,
3026
correct_parents)} for each revision_id where the stored parents
3027
are not correct. dangling_file_versions is a set of (file_id,
3028
revision_id) tuples for versions that are present in this versioned
3029
file, but not used by the corresponding inventory.
3032
self.file_ids = set([file_id for file_id, _ in
3033
self.text_index.iterkeys()])
3034
# text keys is now grouped by file_id
3035
n_weaves = len(self.file_ids)
3036
files_in_revisions = {}
3037
revisions_of_files = {}
3038
n_versions = len(self.text_index)
3039
progress_bar.update('loading text store', 0, n_versions)
3040
parent_map = self.repository.texts.get_parent_map(self.text_index)
3041
# On unlistable transports this could well be empty/error...
3042
text_keys = self.repository.texts.keys()
3043
unused_keys = frozenset(text_keys) - set(self.text_index)
3044
for num, key in enumerate(self.text_index.iterkeys()):
3045
if progress_bar is not None:
3046
progress_bar.update('checking text graph', num, n_versions)
3047
correct_parents = self.calculate_file_version_parents(key)
3049
knit_parents = parent_map[key]
3050
except errors.RevisionNotPresent:
3053
if correct_parents != knit_parents:
3054
wrong_parents[key] = (knit_parents, correct_parents)
3055
return wrong_parents, unused_keys
3058
def _old_get_graph(repository, revision_id):
3059
"""DO NOT USE. That is all. I'm serious."""
3060
graph = repository.get_graph()
3061
revision_graph = dict(((key, value) for key, value in
3062
graph.iter_ancestry([revision_id]) if value is not None))
3063
return _strip_NULL_ghosts(revision_graph)
3066
def _strip_NULL_ghosts(revision_graph):
3067
"""Also don't use this. more compatibility code for unmigrated clients."""
3068
# Filter ghosts, and null:
3069
if _mod_revision.NULL_REVISION in revision_graph:
3070
del revision_graph[_mod_revision.NULL_REVISION]
3071
for key, parents in revision_graph.items():
3072
revision_graph[key] = tuple(parent for parent in parents if parent
3074
return revision_graph
added
removed
bzrlib/repository.py
