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,
588
rev.inventory_sha1 = self.inventories.get_sha1s([key])[key]
589
self._add_revision(rev)
591
def _add_revision(self, revision):
592
text = self._serializer.write_revision_to_string(revision)
593
key = (revision.revision_id,)
594
parents = tuple((parent,) for parent in revision.parent_ids)
595
self.revisions.add_lines(key, parents, osutils.split_lines(text))
597
def all_revision_ids(self):
598
"""Returns a list of all the revision ids in the repository.
600
This is deprecated because code should generally work on the graph
601
reachable from a particular revision, and ignore any other revisions
602
that might be present. There is no direct replacement method.
604
if 'evil' in debug.debug_flags:
605
mutter_callsite(2, "all_revision_ids is linear with history.")
606
return self._all_revision_ids()
608
def _all_revision_ids(self):
609
"""Returns a list of all the revision ids in the repository.
611
These are in as much topological order as the underlying store can
614
raise NotImplementedError(self._all_revision_ids)
616
def break_lock(self):
617
"""Break a lock if one is present from another instance.
619
Uses the ui factory to ask for confirmation if the lock may be from
622
self.control_files.break_lock()
625
def _eliminate_revisions_not_present(self, revision_ids):
626
"""Check every revision id in revision_ids to see if we have it.
628
Returns a set of the present revisions.
631
graph = self.get_graph()
632
parent_map = graph.get_parent_map(revision_ids)
633
# The old API returned a list, should this actually be a set?
634
return parent_map.keys()
637
def create(a_bzrdir):
638
"""Construct the current default format repository in a_bzrdir."""
639
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
641
def __init__(self, _format, a_bzrdir, control_files):
642
"""instantiate a Repository.
644
:param _format: The format of the repository on disk.
645
:param a_bzrdir: The BzrDir of the repository.
647
In the future we will have a single api for all stores for
648
getting file texts, inventories and revisions, then
649
this construct will accept instances of those things.
651
super(Repository, self).__init__()
652
self._format = _format
653
# the following are part of the public API for Repository:
654
self.bzrdir = a_bzrdir
655
self.control_files = control_files
656
self._transport = control_files._transport
657
self.base = self._transport.base
659
self._reconcile_does_inventory_gc = True
660
self._reconcile_fixes_text_parents = False
661
self._reconcile_backsup_inventory = True
662
# not right yet - should be more semantically clear ?
664
# TODO: make sure to construct the right store classes, etc, depending
665
# on whether escaping is required.
666
self._warn_if_deprecated()
667
self._write_group = None
670
return '%s(%r)' % (self.__class__.__name__,
673
def has_same_location(self, other):
674
"""Returns a boolean indicating if this repository is at the same
675
location as another repository.
677
This might return False even when two repository objects are accessing
678
the same physical repository via different URLs.
680
if self.__class__ is not other.__class__:
682
return (self._transport.base == other._transport.base)
684
def is_in_write_group(self):
685
"""Return True if there is an open write group.
687
:seealso: start_write_group.
689
return self._write_group is not None
692
return self.control_files.is_locked()
694
def is_write_locked(self):
695
"""Return True if this object is write locked."""
696
return self.is_locked() and self.control_files._lock_mode == 'w'
698
def lock_write(self, token=None):
699
"""Lock this repository for writing.
701
This causes caching within the repository obejct to start accumlating
702
data during reads, and allows a 'write_group' to be obtained. Write
703
groups must be used for actual data insertion.
705
:param token: if this is already locked, then lock_write will fail
706
unless the token matches the existing lock.
707
:returns: a token if this instance supports tokens, otherwise None.
708
:raises TokenLockingNotSupported: when a token is given but this
709
instance doesn't support using token locks.
710
:raises MismatchedToken: if the specified token doesn't match the token
711
of the existing lock.
712
:seealso: start_write_group.
714
A token should be passed in if you know that you have locked the object
715
some other way, and need to synchronise this object's state with that
718
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
720
result = self.control_files.lock_write(token=token)
725
self.control_files.lock_read()
728
def get_physical_lock_status(self):
729
return self.control_files.get_physical_lock_status()
731
def leave_lock_in_place(self):
732
"""Tell this repository not to release the physical lock when this
735
If lock_write doesn't return a token, then this method is not supported.
737
self.control_files.leave_in_place()
739
def dont_leave_lock_in_place(self):
740
"""Tell this repository to release the physical lock when this
741
object is unlocked, even if it didn't originally acquire it.
743
If lock_write doesn't return a token, then this method is not supported.
745
self.control_files.dont_leave_in_place()
748
def gather_stats(self, revid=None, committers=None):
749
"""Gather statistics from a revision id.
751
:param revid: The revision id to gather statistics from, if None, then
752
no revision specific statistics are gathered.
753
:param committers: Optional parameter controlling whether to grab
754
a count of committers from the revision specific statistics.
755
:return: A dictionary of statistics. Currently this contains:
756
committers: The number of committers if requested.
757
firstrev: A tuple with timestamp, timezone for the penultimate left
758
most ancestor of revid, if revid is not the NULL_REVISION.
759
latestrev: A tuple with timestamp, timezone for revid, if revid is
760
not the NULL_REVISION.
761
revisions: The total revision count in the repository.
762
size: An estimate disk size of the repository in bytes.
765
if revid and committers:
766
result['committers'] = 0
767
if revid and revid != _mod_revision.NULL_REVISION:
769
all_committers = set()
770
revisions = self.get_ancestry(revid)
771
# pop the leading None
773
first_revision = None
775
# ignore the revisions in the middle - just grab first and last
776
revisions = revisions[0], revisions[-1]
777
for revision in self.get_revisions(revisions):
778
if not first_revision:
779
first_revision = revision
781
all_committers.add(revision.committer)
782
last_revision = revision
784
result['committers'] = len(all_committers)
785
result['firstrev'] = (first_revision.timestamp,
786
first_revision.timezone)
787
result['latestrev'] = (last_revision.timestamp,
788
last_revision.timezone)
790
# now gather global repository information
791
# XXX: This is available for many repos regardless of listability.
792
if self.bzrdir.root_transport.listable():
793
# XXX: do we want to __define len__() ?
794
# Maybe the versionedfiles object should provide a different
795
# method to get the number of keys.
796
result['revisions'] = len(self.revisions.keys())
800
def find_branches(self, using=False):
801
"""Find branches underneath this repository.
803
This will include branches inside other branches.
805
:param using: If True, list only branches using this repository.
807
if using and not self.is_shared():
809
return [self.bzrdir.open_branch()]
810
except errors.NotBranchError:
812
class Evaluator(object):
815
self.first_call = True
817
def __call__(self, bzrdir):
818
# On the first call, the parameter is always the bzrdir
819
# containing the current repo.
820
if not self.first_call:
822
repository = bzrdir.open_repository()
823
except errors.NoRepositoryPresent:
826
return False, (None, repository)
827
self.first_call = False
829
value = (bzrdir.open_branch(), None)
830
except errors.NotBranchError:
835
for branch, repository in bzrdir.BzrDir.find_bzrdirs(
836
self.bzrdir.root_transport, evaluate=Evaluator()):
837
if branch is not None:
838
branches.append(branch)
839
if not using and repository is not None:
840
branches.extend(repository.find_branches())
844
def search_missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
845
"""Return the revision ids that other has that this does not.
847
These are returned in topological order.
849
revision_id: only return revision ids included by revision_id.
851
return InterRepository.get(other, self).search_missing_revision_ids(
852
revision_id, find_ghosts)
854
@deprecated_method(one_two)
856
def missing_revision_ids(self, other, revision_id=None, find_ghosts=True):
857
"""Return the revision ids that other has that this does not.
859
These are returned in topological order.
861
revision_id: only return revision ids included by revision_id.
863
keys = self.search_missing_revision_ids(
864
other, revision_id, find_ghosts).get_keys()
867
parents = other.get_graph().get_parent_map(keys)
870
return tsort.topo_sort(parents)
874
"""Open the repository rooted at base.
876
For instance, if the repository is at URL/.bzr/repository,
877
Repository.open(URL) -> a Repository instance.
879
control = bzrdir.BzrDir.open(base)
880
return control.open_repository()
882
def copy_content_into(self, destination, revision_id=None):
883
"""Make a complete copy of the content in self into destination.
885
This is a destructive operation! Do not use it on existing
888
return InterRepository.get(self, destination).copy_content(revision_id)
890
def commit_write_group(self):
891
"""Commit the contents accrued within the current write group.
893
:seealso: start_write_group.
895
if self._write_group is not self.get_transaction():
896
# has an unlock or relock occured ?
897
raise errors.BzrError('mismatched lock context %r and '
899
(self.get_transaction(), self._write_group))
900
self._commit_write_group()
901
self._write_group = None
903
def _commit_write_group(self):
904
"""Template method for per-repository write group cleanup.
906
This is called before the write group is considered to be
907
finished and should ensure that all data handed to the repository
908
for writing during the write group is safely committed (to the
909
extent possible considering file system caching etc).
912
def fetch(self, source, revision_id=None, pb=None, find_ghosts=False):
913
"""Fetch the content required to construct revision_id from source.
915
If revision_id is None all content is copied.
916
:param find_ghosts: Find and copy revisions in the source that are
917
ghosts in the target (and not reachable directly by walking out to
918
the first-present revision in target from revision_id).
920
# fast path same-url fetch operations
921
if self.has_same_location(source):
922
# check that last_revision is in 'from' and then return a
924
if (revision_id is not None and
925
not _mod_revision.is_null(revision_id)):
926
self.get_revision(revision_id)
928
inter = InterRepository.get(source, self)
930
return inter.fetch(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts)
931
except NotImplementedError:
932
raise errors.IncompatibleRepositories(source, self)
934
def create_bundle(self, target, base, fileobj, format=None):
935
return serializer.write_bundle(self, target, base, fileobj, format)
937
def get_commit_builder(self, branch, parents, config, timestamp=None,
938
timezone=None, committer=None, revprops=None,
940
"""Obtain a CommitBuilder for this repository.
942
:param branch: Branch to commit to.
943
:param parents: Revision ids of the parents of the new revision.
944
:param config: Configuration to use.
945
:param timestamp: Optional timestamp recorded for commit.
946
:param timezone: Optional timezone for timestamp.
947
:param committer: Optional committer to set for commit.
948
:param revprops: Optional dictionary of revision properties.
949
:param revision_id: Optional revision id.
951
result = self._commit_builder_class(self, parents, config,
952
timestamp, timezone, committer, revprops, revision_id)
953
self.start_write_group()
957
if (self.control_files._lock_count == 1 and
958
self.control_files._lock_mode == 'w'):
959
if self._write_group is not None:
960
self.abort_write_group()
961
self.control_files.unlock()
962
raise errors.BzrError(
963
'Must end write groups before releasing write locks.')
964
self.control_files.unlock()
967
def clone(self, a_bzrdir, revision_id=None):
968
"""Clone this repository into a_bzrdir using the current format.
970
Currently no check is made that the format of this repository and
971
the bzrdir format are compatible. FIXME RBC 20060201.
973
:return: The newly created destination repository.
975
# TODO: deprecate after 0.16; cloning this with all its settings is
976
# probably not very useful -- mbp 20070423
977
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
978
self.copy_content_into(dest_repo, revision_id)
981
def start_write_group(self):
982
"""Start a write group in the repository.
984
Write groups are used by repositories which do not have a 1:1 mapping
985
between file ids and backend store to manage the insertion of data from
986
both fetch and commit operations.
988
A write lock is required around the start_write_group/commit_write_group
989
for the support of lock-requiring repository formats.
991
One can only insert data into a repository inside a write group.
995
if not self.is_write_locked():
996
raise errors.NotWriteLocked(self)
997
if self._write_group:
998
raise errors.BzrError('already in a write group')
999
self._start_write_group()
1000
# so we can detect unlock/relock - the write group is now entered.
1001
self._write_group = self.get_transaction()
1003
def _start_write_group(self):
1004
"""Template method for per-repository write group startup.
1006
This is called before the write group is considered to be
1011
def sprout(self, to_bzrdir, revision_id=None):
1012
"""Create a descendent repository for new development.
1014
Unlike clone, this does not copy the settings of the repository.
1016
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
1017
dest_repo.fetch(self, revision_id=revision_id)
1020
def _create_sprouting_repo(self, a_bzrdir, shared):
1021
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
1022
# use target default format.
1023
dest_repo = a_bzrdir.create_repository()
1025
# Most control formats need the repository to be specifically
1026
# created, but on some old all-in-one formats it's not needed
1028
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
1029
except errors.UninitializableFormat:
1030
dest_repo = a_bzrdir.open_repository()
1034
def has_revision(self, revision_id):
1035
"""True if this repository has a copy of the revision."""
1036
return revision_id in self.has_revisions((revision_id,))
1039
def has_revisions(self, revision_ids):
1040
"""Probe to find out the presence of multiple revisions.
1042
:param revision_ids: An iterable of revision_ids.
1043
:return: A set of the revision_ids that were present.
1045
parent_map = self.revisions.get_parent_map(
1046
[(rev_id,) for rev_id in revision_ids])
1048
if _mod_revision.NULL_REVISION in revision_ids:
1049
result.add(_mod_revision.NULL_REVISION)
1050
result.update([key[0] for key in parent_map])
1054
def get_revision(self, revision_id):
1055
"""Return the Revision object for a named revision."""
1056
return self.get_revisions([revision_id])[0]
1059
def get_revision_reconcile(self, revision_id):
1060
"""'reconcile' helper routine that allows access to a revision always.
1062
This variant of get_revision does not cross check the weave graph
1063
against the revision one as get_revision does: but it should only
1064
be used by reconcile, or reconcile-alike commands that are correcting
1065
or testing the revision graph.
1067
return self._get_revisions([revision_id])[0]
1070
def get_revisions(self, revision_ids):
1071
"""Get many revisions at once."""
1072
return self._get_revisions(revision_ids)
1075
def _get_revisions(self, revision_ids):
1076
"""Core work logic to get many revisions without sanity checks."""
1077
for rev_id in revision_ids:
1078
if not rev_id or not isinstance(rev_id, basestring):
1079
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
1080
keys = [(key,) for key in revision_ids]
1081
stream = self.revisions.get_record_stream(keys, 'unordered', True)
1083
for record in stream:
1084
if record.storage_kind == 'absent':
1085
raise errors.NoSuchRevision(self, record.key[0])
1086
text = record.get_bytes_as('fulltext')
1087
rev = self._serializer.read_revision_from_string(text)
1088
revs[record.key[0]] = rev
1089
return [revs[revid] for revid in revision_ids]
1092
def get_revision_xml(self, revision_id):
1093
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1094
# would have already do it.
1095
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1096
rev = self.get_revision(revision_id)
1097
rev_tmp = StringIO()
1098
# the current serializer..
1099
self._serializer.write_revision(rev, rev_tmp)
1101
return rev_tmp.getvalue()
1103
def get_deltas_for_revisions(self, revisions):
1104
"""Produce a generator of revision deltas.
1106
Note that the input is a sequence of REVISIONS, not revision_ids.
1107
Trees will be held in memory until the generator exits.
1108
Each delta is relative to the revision's lefthand predecessor.
1110
required_trees = set()
1111
for revision in revisions:
1112
required_trees.add(revision.revision_id)
1113
required_trees.update(revision.parent_ids[:1])
1114
trees = dict((t.get_revision_id(), t) for
1115
t in self.revision_trees(required_trees))
1116
for revision in revisions:
1117
if not revision.parent_ids:
1118
old_tree = self.revision_tree(None)
1120
old_tree = trees[revision.parent_ids[0]]
1121
yield trees[revision.revision_id].changes_from(old_tree)
1124
def get_revision_delta(self, revision_id):
1125
"""Return the delta for one revision.
1127
The delta is relative to the left-hand predecessor of the
1130
r = self.get_revision(revision_id)
1131
return list(self.get_deltas_for_revisions([r]))[0]
1134
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1135
signature = gpg_strategy.sign(plaintext)
1136
self.add_signature_text(revision_id, signature)
1139
def add_signature_text(self, revision_id, signature):
1140
self.signatures.add_lines((revision_id,), (),
1141
osutils.split_lines(signature))
1143
def find_text_key_references(self):
1144
"""Find the text key references within the repository.
1146
:return: a dictionary mapping (file_id, revision_id) tuples to altered file-ids to an iterable of
1147
revision_ids. Each altered file-ids has the exact revision_ids that
1148
altered it listed explicitly.
1149
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1150
to whether they were referred to by the inventory of the
1151
revision_id that they contain. The inventory texts from all present
1152
revision ids are assessed to generate this report.
1154
revision_keys = self.revisions.keys()
1155
w = self.inventories
1156
pb = ui.ui_factory.nested_progress_bar()
1158
return self._find_text_key_references_from_xml_inventory_lines(
1159
w.iter_lines_added_or_present_in_keys(revision_keys, pb=pb))
1163
def _find_text_key_references_from_xml_inventory_lines(self,
1165
"""Core routine for extracting references to texts from inventories.
1167
This performs the translation of xml lines to revision ids.
1169
:param line_iterator: An iterator of lines, origin_version_id
1170
:return: A dictionary mapping text keys ((fileid, revision_id) tuples)
1171
to whether they were referred to by the inventory of the
1172
revision_id that they contain. Note that if that revision_id was
1173
not part of the line_iterator's output then False will be given -
1174
even though it may actually refer to that key.
1176
if not self._serializer.support_altered_by_hack:
1177
raise AssertionError(
1178
"_find_text_key_references_from_xml_inventory_lines only "
1179
"supported for branches which store inventory as unnested xml"
1180
", not on %r" % self)
1183
# this code needs to read every new line in every inventory for the
1184
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1185
# not present in one of those inventories is unnecessary but not
1186
# harmful because we are filtering by the revision id marker in the
1187
# inventory lines : we only select file ids altered in one of those
1188
# revisions. We don't need to see all lines in the inventory because
1189
# only those added in an inventory in rev X can contain a revision=X
1191
unescape_revid_cache = {}
1192
unescape_fileid_cache = {}
1194
# jam 20061218 In a big fetch, this handles hundreds of thousands
1195
# of lines, so it has had a lot of inlining and optimizing done.
1196
# Sorry that it is a little bit messy.
1197
# Move several functions to be local variables, since this is a long
1199
search = self._file_ids_altered_regex.search
1200
unescape = _unescape_xml
1201
setdefault = result.setdefault
1202
for line, line_key in line_iterator:
1203
match = search(line)
1206
# One call to match.group() returning multiple items is quite a
1207
# bit faster than 2 calls to match.group() each returning 1
1208
file_id, revision_id = match.group('file_id', 'revision_id')
1210
# Inlining the cache lookups helps a lot when you make 170,000
1211
# lines and 350k ids, versus 8.4 unique ids.
1212
# Using a cache helps in 2 ways:
1213
# 1) Avoids unnecessary decoding calls
1214
# 2) Re-uses cached strings, which helps in future set and
1216
# (2) is enough that removing encoding entirely along with
1217
# the cache (so we are using plain strings) results in no
1218
# performance improvement.
1220
revision_id = unescape_revid_cache[revision_id]
1222
unescaped = unescape(revision_id)
1223
unescape_revid_cache[revision_id] = unescaped
1224
revision_id = unescaped
1226
# Note that unconditionally unescaping means that we deserialise
1227
# every fileid, which for general 'pull' is not great, but we don't
1228
# really want to have some many fulltexts that this matters anyway.
1231
file_id = unescape_fileid_cache[file_id]
1233
unescaped = unescape(file_id)
1234
unescape_fileid_cache[file_id] = unescaped
1237
key = (file_id, revision_id)
1238
setdefault(key, False)
1239
if revision_id == line_key[-1]:
1243
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1245
"""Helper routine for fileids_altered_by_revision_ids.
1247
This performs the translation of xml lines to revision ids.
1249
:param line_iterator: An iterator of lines, origin_version_id
1250
:param revision_ids: The revision ids to filter for. This should be a
1251
set or other type which supports efficient __contains__ lookups, as
1252
the revision id from each parsed line will be looked up in the
1253
revision_ids filter.
1254
:return: a dictionary mapping altered file-ids to an iterable of
1255
revision_ids. Each altered file-ids has the exact revision_ids that
1256
altered it listed explicitly.
1259
setdefault = result.setdefault
1261
self._find_text_key_references_from_xml_inventory_lines(
1262
line_iterator).iterkeys():
1263
# once data is all ensured-consistent; then this is
1264
# if revision_id == version_id
1265
if key[-1:] in revision_ids:
1266
setdefault(key[0], set()).add(key[-1])
1269
def fileids_altered_by_revision_ids(self, revision_ids, _inv_weave=None):
1270
"""Find the file ids and versions affected by revisions.
1272
:param revisions: an iterable containing revision ids.
1273
:param _inv_weave: The inventory weave from this repository or None.
1274
If None, the inventory weave will be opened automatically.
1275
:return: a dictionary mapping altered file-ids to an iterable of
1276
revision_ids. Each altered file-ids has the exact revision_ids that
1277
altered it listed explicitly.
1279
selected_keys = set((revid,) for revid in revision_ids)
1280
w = _inv_weave or self.inventories
1281
pb = ui.ui_factory.nested_progress_bar()
1283
return self._find_file_ids_from_xml_inventory_lines(
1284
w.iter_lines_added_or_present_in_keys(
1285
selected_keys, pb=pb),
1290
def iter_files_bytes(self, desired_files):
1291
"""Iterate through file versions.
1293
Files will not necessarily be returned in the order they occur in
1294
desired_files. No specific order is guaranteed.
1296
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1297
value supplied by the caller as part of desired_files. It should
1298
uniquely identify the file version in the caller's context. (Examples:
1299
an index number or a TreeTransform trans_id.)
1301
bytes_iterator is an iterable of bytestrings for the file. The
1302
kind of iterable and length of the bytestrings are unspecified, but for
1303
this implementation, it is a list of bytes produced by
1304
VersionedFile.get_record_stream().
1306
:param desired_files: a list of (file_id, revision_id, identifier)
1309
transaction = self.get_transaction()
1311
for file_id, revision_id, callable_data in desired_files:
1312
text_keys[(file_id, revision_id)] = callable_data
1313
for record in self.texts.get_record_stream(text_keys, 'unordered', True):
1314
if record.storage_kind == 'absent':
1315
raise errors.RevisionNotPresent(record.key, self)
1316
yield text_keys[record.key], record.get_bytes_as('fulltext')
1318
def _generate_text_key_index(self, text_key_references=None,
1320
"""Generate a new text key index for the repository.
1322
This is an expensive function that will take considerable time to run.
1324
:return: A dict mapping text keys ((file_id, revision_id) tuples) to a
1325
list of parents, also text keys. When a given key has no parents,
1326
the parents list will be [NULL_REVISION].
1328
# All revisions, to find inventory parents.
1329
if ancestors is None:
1330
graph = self.get_graph()
1331
ancestors = graph.get_parent_map(self.all_revision_ids())
1332
if text_key_references is None:
1333
text_key_references = self.find_text_key_references()
1334
pb = ui.ui_factory.nested_progress_bar()
1336
return self._do_generate_text_key_index(ancestors,
1337
text_key_references, pb)
1341
def _do_generate_text_key_index(self, ancestors, text_key_references, pb):
1342
"""Helper for _generate_text_key_index to avoid deep nesting."""
1343
revision_order = tsort.topo_sort(ancestors)
1344
invalid_keys = set()
1346
for revision_id in revision_order:
1347
revision_keys[revision_id] = set()
1348
text_count = len(text_key_references)
1349
# a cache of the text keys to allow reuse; costs a dict of all the
1350
# keys, but saves a 2-tuple for every child of a given key.
1352
for text_key, valid in text_key_references.iteritems():
1354
invalid_keys.add(text_key)
1356
revision_keys[text_key[1]].add(text_key)
1357
text_key_cache[text_key] = text_key
1358
del text_key_references
1360
text_graph = graph.Graph(graph.DictParentsProvider(text_index))
1361
NULL_REVISION = _mod_revision.NULL_REVISION
1362
# Set a cache with a size of 10 - this suffices for bzr.dev but may be
1363
# too small for large or very branchy trees. However, for 55K path
1364
# trees, it would be easy to use too much memory trivially. Ideally we
1365
# could gauge this by looking at available real memory etc, but this is
1366
# always a tricky proposition.
1367
inventory_cache = lru_cache.LRUCache(10)
1368
batch_size = 10 # should be ~150MB on a 55K path tree
1369
batch_count = len(revision_order) / batch_size + 1
1371
pb.update("Calculating text parents.", processed_texts, text_count)
1372
for offset in xrange(batch_count):
1373
to_query = revision_order[offset * batch_size:(offset + 1) *
1377
for rev_tree in self.revision_trees(to_query):
1378
revision_id = rev_tree.get_revision_id()
1379
parent_ids = ancestors[revision_id]
1380
for text_key in revision_keys[revision_id]:
1381
pb.update("Calculating text parents.", processed_texts)
1382
processed_texts += 1
1383
candidate_parents = []
1384
for parent_id in parent_ids:
1385
parent_text_key = (text_key[0], parent_id)
1387
check_parent = parent_text_key not in \
1388
revision_keys[parent_id]
1390
# the parent parent_id is a ghost:
1391
check_parent = False
1392
# truncate the derived graph against this ghost.
1393
parent_text_key = None
1395
# look at the parent commit details inventories to
1396
# determine possible candidates in the per file graph.
1399
inv = inventory_cache[parent_id]
1401
inv = self.revision_tree(parent_id).inventory
1402
inventory_cache[parent_id] = inv
1403
parent_entry = inv._byid.get(text_key[0], None)
1404
if parent_entry is not None:
1406
text_key[0], parent_entry.revision)
1408
parent_text_key = None
1409
if parent_text_key is not None:
1410
candidate_parents.append(
1411
text_key_cache[parent_text_key])
1412
parent_heads = text_graph.heads(candidate_parents)
1413
new_parents = list(parent_heads)
1414
new_parents.sort(key=lambda x:candidate_parents.index(x))
1415
if new_parents == []:
1416
new_parents = [NULL_REVISION]
1417
text_index[text_key] = new_parents
1419
for text_key in invalid_keys:
1420
text_index[text_key] = [NULL_REVISION]
1423
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1424
"""Get an iterable listing the keys of all the data introduced by a set
1427
The keys will be ordered so that the corresponding items can be safely
1428
fetched and inserted in that order.
1430
:returns: An iterable producing tuples of (knit-kind, file-id,
1431
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1432
'revisions'. file-id is None unless knit-kind is 'file'.
1434
# XXX: it's a bit weird to control the inventory weave caching in this
1435
# generator. Ideally the caching would be done in fetch.py I think. Or
1436
# maybe this generator should explicitly have the contract that it
1437
# should not be iterated until the previously yielded item has been
1439
inv_w = self.inventories
1441
# file ids that changed
1442
file_ids = self.fileids_altered_by_revision_ids(revision_ids, inv_w)
1444
num_file_ids = len(file_ids)
1445
for file_id, altered_versions in file_ids.iteritems():
1446
if _files_pb is not None:
1447
_files_pb.update("fetch texts", count, num_file_ids)
1449
yield ("file", file_id, altered_versions)
1450
# We're done with the files_pb. Note that it finished by the caller,
1451
# just as it was created by the caller.
1455
yield ("inventory", None, revision_ids)
1458
revisions_with_signatures = set()
1459
for rev_id in revision_ids:
1461
self.get_signature_text(rev_id)
1462
except errors.NoSuchRevision:
1466
revisions_with_signatures.add(rev_id)
1467
yield ("signatures", None, revisions_with_signatures)
1470
yield ("revisions", None, revision_ids)
1473
def get_inventory(self, revision_id):
1474
"""Get Inventory object by revision id."""
1475
return self.iter_inventories([revision_id]).next()
1477
def iter_inventories(self, revision_ids):
1478
"""Get many inventories by revision_ids.
1480
This will buffer some or all of the texts used in constructing the
1481
inventories in memory, but will only parse a single inventory at a
1484
:return: An iterator of inventories.
1486
if ((None in revision_ids)
1487
or (_mod_revision.NULL_REVISION in revision_ids)):
1488
raise ValueError('cannot get null revision inventory')
1489
return self._iter_inventories(revision_ids)
1491
def _iter_inventories(self, revision_ids):
1492
"""single-document based inventory iteration."""
1493
for text, revision_id in self._iter_inventory_xmls(revision_ids):
1494
yield self.deserialise_inventory(revision_id, text)
1496
def _iter_inventory_xmls(self, revision_ids):
1497
keys = [(revision_id,) for revision_id in revision_ids]
1498
stream = self.inventories.get_record_stream(keys, 'unordered', True)
1500
for record in stream:
1501
if record.storage_kind != 'absent':
1502
texts[record.key] = record.get_bytes_as('fulltext')
1504
raise errors.NoSuchRevision(self, record.key)
1506
yield texts[key], key[-1]
1508
def deserialise_inventory(self, revision_id, xml):
1509
"""Transform the xml into an inventory object.
1511
:param revision_id: The expected revision id of the inventory.
1512
:param xml: A serialised inventory.
1514
result = self._serializer.read_inventory_from_string(xml, revision_id)
1515
if result.revision_id != revision_id:
1516
raise AssertionError('revision id mismatch %s != %s' % (
1517
result.revision_id, revision_id))
1520
def serialise_inventory(self, inv):
1521
return self._serializer.write_inventory_to_string(inv)
1523
def _serialise_inventory_to_lines(self, inv):
1524
return self._serializer.write_inventory_to_lines(inv)
1526
def get_serializer_format(self):
1527
return self._serializer.format_num
1530
def get_inventory_xml(self, revision_id):
1531
"""Get inventory XML as a file object."""
1532
texts = self._iter_inventory_xmls([revision_id])
1534
text, revision_id = texts.next()
1535
except StopIteration:
1536
raise errors.HistoryMissing(self, 'inventory', revision_id)
1540
def get_inventory_sha1(self, revision_id):
1541
"""Return the sha1 hash of the inventory entry
1543
return self.get_revision(revision_id).inventory_sha1
1545
def iter_reverse_revision_history(self, revision_id):
1546
"""Iterate backwards through revision ids in the lefthand history
1548
:param revision_id: The revision id to start with. All its lefthand
1549
ancestors will be traversed.
1551
graph = self.get_graph()
1552
next_id = revision_id
1554
if next_id in (None, _mod_revision.NULL_REVISION):
1557
# Note: The following line may raise KeyError in the event of
1558
# truncated history. We decided not to have a try:except:raise
1559
# RevisionNotPresent here until we see a use for it, because of the
1560
# cost in an inner loop that is by its very nature O(history).
1561
# Robert Collins 20080326
1562
parents = graph.get_parent_map([next_id])[next_id]
1563
if len(parents) == 0:
1566
next_id = parents[0]
1569
def get_revision_inventory(self, revision_id):
1570
"""Return inventory of a past revision."""
1571
# TODO: Unify this with get_inventory()
1572
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1573
# must be the same as its revision, so this is trivial.
1574
if revision_id is None:
1575
# This does not make sense: if there is no revision,
1576
# then it is the current tree inventory surely ?!
1577
# and thus get_root_id() is something that looks at the last
1578
# commit on the branch, and the get_root_id is an inventory check.
1579
raise NotImplementedError
1580
# return Inventory(self.get_root_id())
1582
return self.get_inventory(revision_id)
1585
def is_shared(self):
1586
"""Return True if this repository is flagged as a shared repository."""
1587
raise NotImplementedError(self.is_shared)
1590
def reconcile(self, other=None, thorough=False):
1591
"""Reconcile this repository."""
1592
from bzrlib.reconcile import RepoReconciler
1593
reconciler = RepoReconciler(self, thorough=thorough)
1594
reconciler.reconcile()
1597
def _refresh_data(self):
1598
"""Helper called from lock_* to ensure coherency with disk.
1600
The default implementation does nothing; it is however possible
1601
for repositories to maintain loaded indices across multiple locks
1602
by checking inside their implementation of this method to see
1603
whether their indices are still valid. This depends of course on
1604
the disk format being validatable in this manner.
1608
def revision_tree(self, revision_id):
1609
"""Return Tree for a revision on this branch.
1611
`revision_id` may be None for the empty tree revision.
1613
# TODO: refactor this to use an existing revision object
1614
# so we don't need to read it in twice.
1615
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1616
return RevisionTree(self, Inventory(root_id=None),
1617
_mod_revision.NULL_REVISION)
1619
inv = self.get_revision_inventory(revision_id)
1620
return RevisionTree(self, inv, revision_id)
1622
def revision_trees(self, revision_ids):
1623
"""Return Tree for a revision on this branch.
1625
`revision_id` may not be None or 'null:'"""
1626
inventories = self.iter_inventories(revision_ids)
1627
for inv in inventories:
1628
yield RevisionTree(self, inv, inv.revision_id)
1631
def get_ancestry(self, revision_id, topo_sorted=True):
1632
"""Return a list of revision-ids integrated by a revision.
1634
The first element of the list is always None, indicating the origin
1635
revision. This might change when we have history horizons, or
1636
perhaps we should have a new API.
1638
This is topologically sorted.
1640
if _mod_revision.is_null(revision_id):
1642
if not self.has_revision(revision_id):
1643
raise errors.NoSuchRevision(self, revision_id)
1644
graph = self.get_graph()
1646
search = graph._make_breadth_first_searcher([revision_id])
1649
found, ghosts = search.next_with_ghosts()
1650
except StopIteration:
1653
if _mod_revision.NULL_REVISION in keys:
1654
keys.remove(_mod_revision.NULL_REVISION)
1656
parent_map = graph.get_parent_map(keys)
1657
keys = tsort.topo_sort(parent_map)
1658
return [None] + list(keys)
1661
"""Compress the data within the repository.
1663
This operation only makes sense for some repository types. For other
1664
types it should be a no-op that just returns.
1666
This stub method does not require a lock, but subclasses should use
1667
@needs_write_lock as this is a long running call its reasonable to
1668
implicitly lock for the user.
1672
@deprecated_method(one_six)
1673
def print_file(self, file, revision_id):
1674
"""Print `file` to stdout.
1676
FIXME RBC 20060125 as John Meinel points out this is a bad api
1677
- it writes to stdout, it assumes that that is valid etc. Fix
1678
by creating a new more flexible convenience function.
1680
tree = self.revision_tree(revision_id)
1681
# use inventory as it was in that revision
1682
file_id = tree.inventory.path2id(file)
1684
# TODO: jam 20060427 Write a test for this code path
1685
# it had a bug in it, and was raising the wrong
1687
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1688
tree.print_file(file_id)
1690
def get_transaction(self):
1691
return self.control_files.get_transaction()
1693
@deprecated_method(one_one)
1694
def get_parents(self, revision_ids):
1695
"""See StackedParentsProvider.get_parents"""
1696
parent_map = self.get_parent_map(revision_ids)
1697
return [parent_map.get(r, None) for r in revision_ids]
1699
def get_parent_map(self, keys):
1700
"""See graph._StackedParentsProvider.get_parent_map"""
1702
for revision_id in keys:
1703
if revision_id is None:
1704
raise ValueError('get_parent_map(None) is not valid')
1705
if revision_id == _mod_revision.NULL_REVISION:
1706
parent_map[revision_id] = ()
1709
parent_id_list = self.get_revision(revision_id).parent_ids
1710
except errors.NoSuchRevision:
1713
if len(parent_id_list) == 0:
1714
parent_ids = (_mod_revision.NULL_REVISION,)
1716
parent_ids = tuple(parent_id_list)
1717
parent_map[revision_id] = parent_ids
1720
def _make_parents_provider(self):
1723
def get_graph(self, other_repository=None):
1724
"""Return the graph walker for this repository format"""
1725
parents_provider = self._make_parents_provider()
1726
if (other_repository is not None and
1727
not self.has_same_location(other_repository)):
1728
parents_provider = graph._StackedParentsProvider(
1729
[parents_provider, other_repository._make_parents_provider()])
1730
return graph.Graph(parents_provider)
1732
def _get_versioned_file_checker(self):
1733
"""Return an object suitable for checking versioned files."""
1734
return _VersionedFileChecker(self)
1736
def revision_ids_to_search_result(self, result_set):
1737
"""Convert a set of revision ids to a graph SearchResult."""
1738
result_parents = set()
1739
for parents in self.get_graph().get_parent_map(
1740
result_set).itervalues():
1741
result_parents.update(parents)
1742
included_keys = result_set.intersection(result_parents)
1743
start_keys = result_set.difference(included_keys)
1744
exclude_keys = result_parents.difference(result_set)
1745
result = graph.SearchResult(start_keys, exclude_keys,
1746
len(result_set), result_set)
1750
def set_make_working_trees(self, new_value):
1751
"""Set the policy flag for making working trees when creating branches.
1753
This only applies to branches that use this repository.
1755
The default is 'True'.
1756
:param new_value: True to restore the default, False to disable making
1759
raise NotImplementedError(self.set_make_working_trees)
1761
def make_working_trees(self):
1762
"""Returns the policy for making working trees on new branches."""
1763
raise NotImplementedError(self.make_working_trees)
1766
def sign_revision(self, revision_id, gpg_strategy):
1767
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1768
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1771
def has_signature_for_revision_id(self, revision_id):
1772
"""Query for a revision signature for revision_id in the repository."""
1773
if not self.has_revision(revision_id):
1774
raise errors.NoSuchRevision(self, revision_id)
1775
sig_present = (1 == len(
1776
self.signatures.get_parent_map([(revision_id,)])))
1780
def get_signature_text(self, revision_id):
1781
"""Return the text for a signature."""
1782
stream = self.signatures.get_record_stream([(revision_id,)],
1784
record = stream.next()
1785
if record.storage_kind == 'absent':
1786
raise errors.NoSuchRevision(self, revision_id)
1787
return record.get_bytes_as('fulltext')
1790
def check(self, revision_ids=None):
1791
"""Check consistency of all history of given revision_ids.
1793
Different repository implementations should override _check().
1795
:param revision_ids: A non-empty list of revision_ids whose ancestry
1796
will be checked. Typically the last revision_id of a branch.
1798
return self._check(revision_ids)
1800
def _check(self, revision_ids):
1801
result = check.Check(self)
1805
def _warn_if_deprecated(self):
1806
global _deprecation_warning_done
1807
if _deprecation_warning_done:
1809
_deprecation_warning_done = True
1810
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1811
% (self._format, self.bzrdir.transport.base))
1813
def supports_rich_root(self):
1814
return self._format.rich_root_data
1816
def _check_ascii_revisionid(self, revision_id, method):
1817
"""Private helper for ascii-only repositories."""
1818
# weave repositories refuse to store revisionids that are non-ascii.
1819
if revision_id is not None:
1820
# weaves require ascii revision ids.
1821
if isinstance(revision_id, unicode):
1823
revision_id.encode('ascii')
1824
except UnicodeEncodeError:
1825
raise errors.NonAsciiRevisionId(method, self)
1828
revision_id.decode('ascii')
1829
except UnicodeDecodeError:
1830
raise errors.NonAsciiRevisionId(method, self)
1832
def revision_graph_can_have_wrong_parents(self):
1833
"""Is it possible for this repository to have a revision graph with
1836
If True, then this repository must also implement
1837
_find_inconsistent_revision_parents so that check and reconcile can
1838
check for inconsistencies before proceeding with other checks that may
1839
depend on the revision index being consistent.
1841
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1844
# remove these delegates a while after bzr 0.15
1845
def __make_delegated(name, from_module):
1846
def _deprecated_repository_forwarder():
1847
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1848
% (name, from_module),
1851
m = __import__(from_module, globals(), locals(), [name])
1853
return getattr(m, name)
1854
except AttributeError:
1855
raise AttributeError('module %s has no name %s'
1857
globals()[name] = _deprecated_repository_forwarder
1860
'AllInOneRepository',
1861
'WeaveMetaDirRepository',
1862
'PreSplitOutRepositoryFormat',
1863
'RepositoryFormat4',
1864
'RepositoryFormat5',
1865
'RepositoryFormat6',
1866
'RepositoryFormat7',
1868
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1872
'RepositoryFormatKnit',
1873
'RepositoryFormatKnit1',
1875
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1878
def install_revision(repository, rev, revision_tree):
1879
"""Install all revision data into a repository."""
1880
install_revisions(repository, [(rev, revision_tree, None)])
1883
def install_revisions(repository, iterable, num_revisions=None, pb=None):
1884
"""Install all revision data into a repository.
1886
Accepts an iterable of revision, tree, signature tuples. The signature
1889
repository.start_write_group()
1891
for n, (revision, revision_tree, signature) in enumerate(iterable):
1892
_install_revision(repository, revision, revision_tree, signature)
1894
pb.update('Transferring revisions', n + 1, num_revisions)
1896
repository.abort_write_group()
1899
repository.commit_write_group()
1902
def _install_revision(repository, rev, revision_tree, signature):
1903
"""Install all revision data into a repository."""
1904
present_parents = []
1906
for p_id in rev.parent_ids:
1907
if repository.has_revision(p_id):
1908
present_parents.append(p_id)
1909
parent_trees[p_id] = repository.revision_tree(p_id)
1911
parent_trees[p_id] = repository.revision_tree(None)
1913
inv = revision_tree.inventory
1914
entries = inv.iter_entries()
1915
# backwards compatibility hack: skip the root id.
1916
if not repository.supports_rich_root():
1917
path, root = entries.next()
1918
if root.revision != rev.revision_id:
1919
raise errors.IncompatibleRevision(repr(repository))
1921
for path, ie in entries:
1922
text_keys[(ie.file_id, ie.revision)] = ie
1923
text_parent_map = repository.texts.get_parent_map(text_keys)
1924
missing_texts = set(text_keys) - set(text_parent_map)
1925
# Add the texts that are not already present
1926
for text_key in missing_texts:
1927
ie = text_keys[text_key]
1929
# FIXME: TODO: The following loop overlaps/duplicates that done by
1930
# commit to determine parents. There is a latent/real bug here where
1931
# the parents inserted are not those commit would do - in particular
1932
# they are not filtered by heads(). RBC, AB
1933
for revision, tree in parent_trees.iteritems():
1934
if ie.file_id not in tree:
1936
parent_id = tree.inventory[ie.file_id].revision
1937
if parent_id in text_parents:
1939
text_parents.append((ie.file_id, parent_id))
1940
lines = revision_tree.get_file(ie.file_id).readlines()
1941
repository.texts.add_lines(text_key, text_parents, lines)
1943
# install the inventory
1944
repository.add_inventory(rev.revision_id, inv, present_parents)
1945
except errors.RevisionAlreadyPresent:
1947
if signature is not None:
1948
repository.add_signature_text(rev.revision_id, signature)
1949
repository.add_revision(rev.revision_id, rev, inv)
1952
class MetaDirRepository(Repository):
1953
"""Repositories in the new meta-dir layout.
1955
:ivar _transport: Transport for access to repository control files,
1956
typically pointing to .bzr/repository.
1959
def __init__(self, _format, a_bzrdir, control_files):
1960
super(MetaDirRepository, self).__init__(_format, a_bzrdir, control_files)
1961
self._transport = control_files._transport
1964
def is_shared(self):
1965
"""Return True if this repository is flagged as a shared repository."""
1966
return self._transport.has('shared-storage')
1969
def set_make_working_trees(self, new_value):
1970
"""Set the policy flag for making working trees when creating branches.
1972
This only applies to branches that use this repository.
1974
The default is 'True'.
1975
:param new_value: True to restore the default, False to disable making
1980
self._transport.delete('no-working-trees')
1981
except errors.NoSuchFile:
1984
self._transport.put_bytes('no-working-trees', '',
1985
mode=self.bzrdir._get_file_mode())
1987
def make_working_trees(self):
1988
"""Returns the policy for making working trees on new branches."""
1989
return not self._transport.has('no-working-trees')
1992
class MetaDirVersionedFileRepository(MetaDirRepository):
1993
"""Repositories in a meta-dir, that work via versioned file objects."""
1995
def __init__(self, _format, a_bzrdir, control_files):
1996
super(MetaDirVersionedFileRepository, self).__init__(_format, a_bzrdir,
2000
class RepositoryFormatRegistry(registry.Registry):
2001
"""Registry of RepositoryFormats."""
2003
def get(self, format_string):
2004
r = registry.Registry.get(self, format_string)
2010
format_registry = RepositoryFormatRegistry()
2011
"""Registry of formats, indexed by their identifying format string.
2013
This can contain either format instances themselves, or classes/factories that
2014
can be called to obtain one.
2018
#####################################################################
2019
# Repository Formats
2021
class RepositoryFormat(object):
2022
"""A repository format.
2024
Formats provide three things:
2025
* An initialization routine to construct repository data on disk.
2026
* a format string which is used when the BzrDir supports versioned
2028
* an open routine which returns a Repository instance.
2030
There is one and only one Format subclass for each on-disk format. But
2031
there can be one Repository subclass that is used for several different
2032
formats. The _format attribute on a Repository instance can be used to
2033
determine the disk format.
2035
Formats are placed in an dict by their format string for reference
2036
during opening. These should be subclasses of RepositoryFormat
2039
Once a format is deprecated, just deprecate the initialize and open
2040
methods on the format class. Do not deprecate the object, as the
2041
object will be created every system load.
2043
Common instance attributes:
2044
_matchingbzrdir - the bzrdir format that the repository format was
2045
originally written to work with. This can be used if manually
2046
constructing a bzrdir and repository, or more commonly for test suite
2050
# Set to True or False in derived classes. True indicates that the format
2051
# supports ghosts gracefully.
2052
supports_ghosts = None
2053
# Can this repository be given external locations to lookup additional
2054
# data. Set to True or False in derived classes.
2055
supports_external_lookups = None
2058
return "<%s>" % self.__class__.__name__
2060
def __eq__(self, other):
2061
# format objects are generally stateless
2062
return isinstance(other, self.__class__)
2064
def __ne__(self, other):
2065
return not self == other
2068
def find_format(klass, a_bzrdir):
2069
"""Return the format for the repository object in a_bzrdir.
2071
This is used by bzr native formats that have a "format" file in
2072
the repository. Other methods may be used by different types of
2076
transport = a_bzrdir.get_repository_transport(None)
2077
format_string = transport.get("format").read()
2078
return format_registry.get(format_string)
2079
except errors.NoSuchFile:
2080
raise errors.NoRepositoryPresent(a_bzrdir)
2082
raise errors.UnknownFormatError(format=format_string,
2086
def register_format(klass, format):
2087
format_registry.register(format.get_format_string(), format)
2090
def unregister_format(klass, format):
2091
format_registry.remove(format.get_format_string())
2094
def get_default_format(klass):
2095
"""Return the current default format."""
2096
from bzrlib import bzrdir
2097
return bzrdir.format_registry.make_bzrdir('default').repository_format
2099
def get_format_string(self):
2100
"""Return the ASCII format string that identifies this format.
2102
Note that in pre format ?? repositories the format string is
2103
not permitted nor written to disk.
2105
raise NotImplementedError(self.get_format_string)
2107
def get_format_description(self):
2108
"""Return the short description for this format."""
2109
raise NotImplementedError(self.get_format_description)
2111
# TODO: this shouldn't be in the base class, it's specific to things that
2112
# use weaves or knits -- mbp 20070207
2113
def _get_versioned_file_store(self,
2118
versionedfile_class=None,
2119
versionedfile_kwargs={},
2121
if versionedfile_class is None:
2122
versionedfile_class = self._versionedfile_class
2123
weave_transport = control_files._transport.clone(name)
2124
dir_mode = control_files._dir_mode
2125
file_mode = control_files._file_mode
2126
return VersionedFileStore(weave_transport, prefixed=prefixed,
2128
file_mode=file_mode,
2129
versionedfile_class=versionedfile_class,
2130
versionedfile_kwargs=versionedfile_kwargs,
2133
def initialize(self, a_bzrdir, shared=False):
2134
"""Initialize a repository of this format in a_bzrdir.
2136
:param a_bzrdir: The bzrdir to put the new repository in it.
2137
:param shared: The repository should be initialized as a sharable one.
2138
:returns: The new repository object.
2140
This may raise UninitializableFormat if shared repository are not
2141
compatible the a_bzrdir.
2143
raise NotImplementedError(self.initialize)
2145
def is_supported(self):
2146
"""Is this format supported?
2148
Supported formats must be initializable and openable.
2149
Unsupported formats may not support initialization or committing or
2150
some other features depending on the reason for not being supported.
2154
def check_conversion_target(self, target_format):
2155
raise NotImplementedError(self.check_conversion_target)
2157
def open(self, a_bzrdir, _found=False):
2158
"""Return an instance of this format for the bzrdir a_bzrdir.
2160
_found is a private parameter, do not use it.
2162
raise NotImplementedError(self.open)
2165
class MetaDirRepositoryFormat(RepositoryFormat):
2166
"""Common base class for the new repositories using the metadir layout."""
2168
rich_root_data = False
2169
supports_tree_reference = False
2170
supports_external_lookups = False
2171
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
2174
super(MetaDirRepositoryFormat, self).__init__()
2176
def _create_control_files(self, a_bzrdir):
2177
"""Create the required files and the initial control_files object."""
2178
# FIXME: RBC 20060125 don't peek under the covers
2179
# NB: no need to escape relative paths that are url safe.
2180
repository_transport = a_bzrdir.get_repository_transport(self)
2181
control_files = lockable_files.LockableFiles(repository_transport,
2182
'lock', lockdir.LockDir)
2183
control_files.create_lock()
2184
return control_files
2186
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
2187
"""Upload the initial blank content."""
2188
control_files = self._create_control_files(a_bzrdir)
2189
control_files.lock_write()
2190
transport = control_files._transport
2192
utf8_files += [('shared-storage', '')]
2194
transport.mkdir_multi(dirs, mode=a_bzrdir._get_dir_mode())
2195
for (filename, content_stream) in files:
2196
transport.put_file(filename, content_stream,
2197
mode=a_bzrdir._get_file_mode())
2198
for (filename, content_bytes) in utf8_files:
2199
transport.put_bytes_non_atomic(filename, content_bytes,
2200
mode=a_bzrdir._get_file_mode())
2202
control_files.unlock()
2205
# formats which have no format string are not discoverable
2206
# and not independently creatable, so are not registered. They're
2207
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
2208
# needed, it's constructed directly by the BzrDir. Non-native formats where
2209
# the repository is not separately opened are similar.
2211
format_registry.register_lazy(
2212
'Bazaar-NG Repository format 7',
2213
'bzrlib.repofmt.weaverepo',
2217
format_registry.register_lazy(
2218
'Bazaar-NG Knit Repository Format 1',
2219
'bzrlib.repofmt.knitrepo',
2220
'RepositoryFormatKnit1',
2223
format_registry.register_lazy(
2224
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
2225
'bzrlib.repofmt.knitrepo',
2226
'RepositoryFormatKnit3',
2229
format_registry.register_lazy(
2230
'Bazaar Knit Repository Format 4 (bzr 1.0)\n',
2231
'bzrlib.repofmt.knitrepo',
2232
'RepositoryFormatKnit4',
2235
# Pack-based formats. There is one format for pre-subtrees, and one for
2236
# post-subtrees to allow ease of testing.
2237
# NOTE: These are experimental in 0.92. Stable in 1.0 and above
2238
format_registry.register_lazy(
2239
'Bazaar pack repository format 1 (needs bzr 0.92)\n',
2240
'bzrlib.repofmt.pack_repo',
2241
'RepositoryFormatKnitPack1',
2243
format_registry.register_lazy(
2244
'Bazaar pack repository format 1 with subtree support (needs bzr 0.92)\n',
2245
'bzrlib.repofmt.pack_repo',
2246
'RepositoryFormatKnitPack3',
2248
format_registry.register_lazy(
2249
'Bazaar pack repository format 1 with rich root (needs bzr 1.0)\n',
2250
'bzrlib.repofmt.pack_repo',
2251
'RepositoryFormatKnitPack4',
2253
# Development formats.
2255
# development 0 - stub to introduce development versioning scheme.
2256
format_registry.register_lazy(
2257
"Bazaar development format 0 (needs bzr.dev from before 1.3)\n",
2258
'bzrlib.repofmt.pack_repo',
2259
'RepositoryFormatPackDevelopment0',
2261
format_registry.register_lazy(
2262
("Bazaar development format 0 with subtree support "
2263
"(needs bzr.dev from before 1.3)\n"),
2264
'bzrlib.repofmt.pack_repo',
2265
'RepositoryFormatPackDevelopment0Subtree',
2267
# 1.3->1.4 go below here
2270
class InterRepository(InterObject):
2271
"""This class represents operations taking place between two repositories.
2273
Its instances have methods like copy_content and fetch, and contain
2274
references to the source and target repositories these operations can be
2277
Often we will provide convenience methods on 'repository' which carry out
2278
operations with another repository - they will always forward to
2279
InterRepository.get(other).method_name(parameters).
2283
"""The available optimised InterRepository types."""
2285
def copy_content(self, revision_id=None):
2286
raise NotImplementedError(self.copy_content)
2288
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2289
"""Fetch the content required to construct revision_id.
2291
The content is copied from self.source to self.target.
2293
:param revision_id: if None all content is copied, if NULL_REVISION no
2295
:param pb: optional progress bar to use for progress reports. If not
2296
provided a default one will be created.
2298
Returns the copied revision count and the failed revisions in a tuple:
2301
raise NotImplementedError(self.fetch)
2303
def _walk_to_common_revisions(self, revision_ids):
2304
"""Walk out from revision_ids in source to revisions target has.
2306
:param revision_ids: The start point for the search.
2307
:return: A set of revision ids.
2309
target_graph = self.target.get_graph()
2310
revision_ids = frozenset(revision_ids)
2311
if set(target_graph.get_parent_map(revision_ids)) == revision_ids:
2312
return graph.SearchResult(revision_ids, set(), 0, set())
2313
missing_revs = set()
2314
source_graph = self.source.get_graph()
2315
# ensure we don't pay silly lookup costs.
2316
searcher = source_graph._make_breadth_first_searcher(revision_ids)
2317
null_set = frozenset([_mod_revision.NULL_REVISION])
2320
next_revs, ghosts = searcher.next_with_ghosts()
2321
except StopIteration:
2323
if revision_ids.intersection(ghosts):
2324
absent_ids = set(revision_ids.intersection(ghosts))
2325
# If all absent_ids are present in target, no error is needed.
2326
absent_ids.difference_update(
2327
set(target_graph.get_parent_map(absent_ids)))
2329
raise errors.NoSuchRevision(self.source, absent_ids.pop())
2330
# we don't care about other ghosts as we can't fetch them and
2331
# haven't been asked to.
2332
next_revs = set(next_revs)
2333
# we always have NULL_REVISION present.
2334
have_revs = set(target_graph.get_parent_map(next_revs)).union(null_set)
2335
missing_revs.update(next_revs - have_revs)
2336
searcher.stop_searching_any(have_revs)
2337
return searcher.get_result()
2339
@deprecated_method(one_two)
2341
def missing_revision_ids(self, revision_id=None, find_ghosts=True):
2342
"""Return the revision ids that source has that target does not.
2344
These are returned in topological order.
2346
:param revision_id: only return revision ids included by this
2348
:param find_ghosts: If True find missing revisions in deep history
2349
rather than just finding the surface difference.
2351
return list(self.search_missing_revision_ids(
2352
revision_id, find_ghosts).get_keys())
2355
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2356
"""Return the revision ids that source has that target does not.
2358
:param revision_id: only return revision ids included by this
2360
:param find_ghosts: If True find missing revisions in deep history
2361
rather than just finding the surface difference.
2362
:return: A bzrlib.graph.SearchResult.
2364
# stop searching at found target revisions.
2365
if not find_ghosts and revision_id is not None:
2366
return self._walk_to_common_revisions([revision_id])
2367
# generic, possibly worst case, slow code path.
2368
target_ids = set(self.target.all_revision_ids())
2369
if revision_id is not None:
2370
source_ids = self.source.get_ancestry(revision_id)
2371
if source_ids[0] is not None:
2372
raise AssertionError()
2375
source_ids = self.source.all_revision_ids()
2376
result_set = set(source_ids).difference(target_ids)
2377
return self.source.revision_ids_to_search_result(result_set)
2380
def _same_model(source, target):
2381
"""True if source and target have the same data representation."""
2382
if source.supports_rich_root() != target.supports_rich_root():
2384
if source._serializer != target._serializer:
2389
class InterSameDataRepository(InterRepository):
2390
"""Code for converting between repositories that represent the same data.
2392
Data format and model must match for this to work.
2396
def _get_repo_format_to_test(self):
2397
"""Repository format for testing with.
2399
InterSameData can pull from subtree to subtree and from non-subtree to
2400
non-subtree, so we test this with the richest repository format.
2402
from bzrlib.repofmt import knitrepo
2403
return knitrepo.RepositoryFormatKnit3()
2406
def is_compatible(source, target):
2407
return InterRepository._same_model(source, target)
2410
def copy_content(self, revision_id=None):
2411
"""Make a complete copy of the content in self into destination.
2413
This copies both the repository's revision data, and configuration information
2414
such as the make_working_trees setting.
2416
This is a destructive operation! Do not use it on existing
2419
:param revision_id: Only copy the content needed to construct
2420
revision_id and its parents.
2423
self.target.set_make_working_trees(self.source.make_working_trees())
2424
except NotImplementedError:
2426
# but don't bother fetching if we have the needed data now.
2427
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2428
self.target.has_revision(revision_id)):
2430
self.target.fetch(self.source, revision_id=revision_id)
2433
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2434
"""See InterRepository.fetch()."""
2435
from bzrlib.fetch import GenericRepoFetcher
2436
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2437
self.source, self.source._format, self.target,
2438
self.target._format)
2439
f = GenericRepoFetcher(to_repository=self.target,
2440
from_repository=self.source,
2441
last_revision=revision_id,
2442
pb=pb, find_ghosts=find_ghosts)
2443
return f.count_copied, f.failed_revisions
2446
class InterWeaveRepo(InterSameDataRepository):
2447
"""Optimised code paths between Weave based repositories.
2449
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2450
implemented lazy inter-object optimisation.
2454
def _get_repo_format_to_test(self):
2455
from bzrlib.repofmt import weaverepo
2456
return weaverepo.RepositoryFormat7()
2459
def is_compatible(source, target):
2460
"""Be compatible with known Weave formats.
2462
We don't test for the stores being of specific types because that
2463
could lead to confusing results, and there is no need to be
2466
from bzrlib.repofmt.weaverepo import (
2472
return (isinstance(source._format, (RepositoryFormat5,
2474
RepositoryFormat7)) and
2475
isinstance(target._format, (RepositoryFormat5,
2477
RepositoryFormat7)))
2478
except AttributeError:
2482
def copy_content(self, revision_id=None):
2483
"""See InterRepository.copy_content()."""
2484
# weave specific optimised path:
2486
self.target.set_make_working_trees(self.source.make_working_trees())
2487
except (errors.RepositoryUpgradeRequired, NotImplemented):
2489
# FIXME do not peek!
2490
if self.source._transport.listable():
2491
pb = ui.ui_factory.nested_progress_bar()
2493
self.target.texts.insert_record_stream(
2494
self.source.texts.get_record_stream(
2495
self.source.texts.keys(), 'topological', False))
2496
pb.update('copying inventory', 0, 1)
2497
self.target.inventories.insert_record_stream(
2498
self.source.inventories.get_record_stream(
2499
self.source.inventories.keys(), 'topological', False))
2500
self.target.signatures.insert_record_stream(
2501
self.source.signatures.get_record_stream(
2502
self.source.signatures.keys(),
2504
self.target.revisions.insert_record_stream(
2505
self.source.revisions.get_record_stream(
2506
self.source.revisions.keys(),
2507
'topological', True))
2511
self.target.fetch(self.source, revision_id=revision_id)
2514
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2515
"""See InterRepository.fetch()."""
2516
from bzrlib.fetch import GenericRepoFetcher
2517
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2518
self.source, self.source._format, self.target, self.target._format)
2519
f = GenericRepoFetcher(to_repository=self.target,
2520
from_repository=self.source,
2521
last_revision=revision_id,
2522
pb=pb, find_ghosts=find_ghosts)
2523
return f.count_copied, f.failed_revisions
2526
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2527
"""See InterRepository.missing_revision_ids()."""
2528
# we want all revisions to satisfy revision_id in source.
2529
# but we don't want to stat every file here and there.
2530
# we want then, all revisions other needs to satisfy revision_id
2531
# checked, but not those that we have locally.
2532
# so the first thing is to get a subset of the revisions to
2533
# satisfy revision_id in source, and then eliminate those that
2534
# we do already have.
2535
# this is slow on high latency connection to self, but as as this
2536
# disk format scales terribly for push anyway due to rewriting
2537
# inventory.weave, this is considered acceptable.
2539
if revision_id is not None:
2540
source_ids = self.source.get_ancestry(revision_id)
2541
if source_ids[0] is not None:
2542
raise AssertionError()
2545
source_ids = self.source._all_possible_ids()
2546
source_ids_set = set(source_ids)
2547
# source_ids is the worst possible case we may need to pull.
2548
# now we want to filter source_ids against what we actually
2549
# have in target, but don't try to check for existence where we know
2550
# we do not have a revision as that would be pointless.
2551
target_ids = set(self.target._all_possible_ids())
2552
possibly_present_revisions = target_ids.intersection(source_ids_set)
2553
actually_present_revisions = set(
2554
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2555
required_revisions = source_ids_set.difference(actually_present_revisions)
2556
if revision_id is not None:
2557
# we used get_ancestry to determine source_ids then we are assured all
2558
# revisions referenced are present as they are installed in topological order.
2559
# and the tip revision was validated by get_ancestry.
2560
result_set = required_revisions
2562
# if we just grabbed the possibly available ids, then
2563
# we only have an estimate of whats available and need to validate
2564
# that against the revision records.
2566
self.source._eliminate_revisions_not_present(required_revisions))
2567
return self.source.revision_ids_to_search_result(result_set)
2570
class InterKnitRepo(InterSameDataRepository):
2571
"""Optimised code paths between Knit based repositories."""
2574
def _get_repo_format_to_test(self):
2575
from bzrlib.repofmt import knitrepo
2576
return knitrepo.RepositoryFormatKnit1()
2579
def is_compatible(source, target):
2580
"""Be compatible with known Knit formats.
2582
We don't test for the stores being of specific types because that
2583
could lead to confusing results, and there is no need to be
2586
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2588
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2589
isinstance(target._format, RepositoryFormatKnit))
2590
except AttributeError:
2592
return are_knits and InterRepository._same_model(source, target)
2595
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2596
"""See InterRepository.fetch()."""
2597
from bzrlib.fetch import KnitRepoFetcher
2598
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2599
self.source, self.source._format, self.target, self.target._format)
2600
f = KnitRepoFetcher(to_repository=self.target,
2601
from_repository=self.source,
2602
last_revision=revision_id,
2603
pb=pb, find_ghosts=find_ghosts)
2604
return f.count_copied, f.failed_revisions
2607
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2608
"""See InterRepository.missing_revision_ids()."""
2609
if revision_id is not None:
2610
source_ids = self.source.get_ancestry(revision_id)
2611
if source_ids[0] is not None:
2612
raise AssertionError()
2615
source_ids = self.source.all_revision_ids()
2616
source_ids_set = set(source_ids)
2617
# source_ids is the worst possible case we may need to pull.
2618
# now we want to filter source_ids against what we actually
2619
# have in target, but don't try to check for existence where we know
2620
# we do not have a revision as that would be pointless.
2621
target_ids = set(self.target.all_revision_ids())
2622
possibly_present_revisions = target_ids.intersection(source_ids_set)
2623
actually_present_revisions = set(
2624
self.target._eliminate_revisions_not_present(possibly_present_revisions))
2625
required_revisions = source_ids_set.difference(actually_present_revisions)
2626
if revision_id is not None:
2627
# we used get_ancestry to determine source_ids then we are assured all
2628
# revisions referenced are present as they are installed in topological order.
2629
# and the tip revision was validated by get_ancestry.
2630
result_set = required_revisions
2632
# if we just grabbed the possibly available ids, then
2633
# we only have an estimate of whats available and need to validate
2634
# that against the revision records.
2636
self.source._eliminate_revisions_not_present(required_revisions))
2637
return self.source.revision_ids_to_search_result(result_set)
2640
class InterPackRepo(InterSameDataRepository):
2641
"""Optimised code paths between Pack based repositories."""
2644
def _get_repo_format_to_test(self):
2645
from bzrlib.repofmt import pack_repo
2646
return pack_repo.RepositoryFormatKnitPack1()
2649
def is_compatible(source, target):
2650
"""Be compatible with known Pack formats.
2652
We don't test for the stores being of specific types because that
2653
could lead to confusing results, and there is no need to be
2656
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2658
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2659
isinstance(target._format, RepositoryFormatPack))
2660
except AttributeError:
2662
return are_packs and InterRepository._same_model(source, target)
2665
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2666
"""See InterRepository.fetch()."""
2667
from bzrlib.repofmt.pack_repo import Packer
2668
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2669
self.source, self.source._format, self.target, self.target._format)
2670
self.count_copied = 0
2671
if revision_id is None:
2673
# everything to do - use pack logic
2674
# to fetch from all packs to one without
2675
# inventory parsing etc, IFF nothing to be copied is in the target.
2677
revision_ids = self.source.all_revision_ids()
2678
revision_keys = [(revid,) for revid in revision_ids]
2679
index = self.target._pack_collection.revision_index.combined_index
2680
present_revision_ids = set(item[1][0] for item in
2681
index.iter_entries(revision_keys))
2682
revision_ids = set(revision_ids) - present_revision_ids
2683
# implementing the TODO will involve:
2684
# - detecting when all of a pack is selected
2685
# - avoiding as much as possible pre-selection, so the
2686
# more-core routines such as create_pack_from_packs can filter in
2687
# a just-in-time fashion. (though having a HEADS list on a
2688
# repository might make this a lot easier, because we could
2689
# sensibly detect 'new revisions' without doing a full index scan.
2690
elif _mod_revision.is_null(revision_id):
2695
revision_ids = self.search_missing_revision_ids(revision_id,
2696
find_ghosts=find_ghosts).get_keys()
2697
except errors.NoSuchRevision:
2698
raise errors.InstallFailed([revision_id])
2699
if len(revision_ids) == 0:
2701
packs = self.source._pack_collection.all_packs()
2702
pack = Packer(self.target._pack_collection, packs, '.fetch',
2703
revision_ids).pack()
2704
if pack is not None:
2705
self.target._pack_collection._save_pack_names()
2706
# Trigger an autopack. This may duplicate effort as we've just done
2707
# a pack creation, but for now it is simpler to think about as
2708
# 'upload data, then repack if needed'.
2709
self.target._pack_collection.autopack()
2710
return (pack.get_revision_count(), [])
2715
def search_missing_revision_ids(self, revision_id=None, find_ghosts=True):
2716
"""See InterRepository.missing_revision_ids().
2718
:param find_ghosts: Find ghosts throughout the ancestry of
2721
if not find_ghosts and revision_id is not None:
2722
return self._walk_to_common_revisions([revision_id])
2723
elif revision_id is not None:
2724
source_ids = self.source.get_ancestry(revision_id)
2725
if source_ids[0] is not None:
2726
raise AssertionError()
2729
source_ids = self.source.all_revision_ids()
2730
# source_ids is the worst possible case we may need to pull.
2731
# now we want to filter source_ids against what we actually
2732
# have in target, but don't try to check for existence where we know
2733
# we do not have a revision as that would be pointless.
2734
target_ids = set(self.target.all_revision_ids())
2735
result_set = set(source_ids).difference(target_ids)
2736
return self.source.revision_ids_to_search_result(result_set)
2739
class InterModel1and2(InterRepository):
2742
def _get_repo_format_to_test(self):
2746
def is_compatible(source, target):
2747
if not source.supports_rich_root() and target.supports_rich_root():
2753
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2754
"""See InterRepository.fetch()."""
2755
from bzrlib.fetch import Model1toKnit2Fetcher
2756
f = Model1toKnit2Fetcher(to_repository=self.target,
2757
from_repository=self.source,
2758
last_revision=revision_id,
2759
pb=pb, find_ghosts=find_ghosts)
2760
return f.count_copied, f.failed_revisions
2763
def copy_content(self, revision_id=None):
2764
"""Make a complete copy of the content in self into destination.
2766
This is a destructive operation! Do not use it on existing
2769
:param revision_id: Only copy the content needed to construct
2770
revision_id and its parents.
2773
self.target.set_make_working_trees(self.source.make_working_trees())
2774
except NotImplementedError:
2776
# but don't bother fetching if we have the needed data now.
2777
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2778
self.target.has_revision(revision_id)):
2780
self.target.fetch(self.source, revision_id=revision_id)
2783
class InterKnit1and2(InterKnitRepo):
2786
def _get_repo_format_to_test(self):
2790
def is_compatible(source, target):
2791
"""Be compatible with Knit1 source and Knit3 target"""
2792
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2794
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2795
RepositoryFormatKnit3)
2796
from bzrlib.repofmt.pack_repo import (
2797
RepositoryFormatKnitPack1,
2798
RepositoryFormatKnitPack3,
2799
RepositoryFormatPackDevelopment0,
2800
RepositoryFormatPackDevelopment0Subtree,
2803
RepositoryFormatKnit1,
2804
RepositoryFormatKnitPack1,
2805
RepositoryFormatPackDevelopment0,
2808
RepositoryFormatKnit3,
2809
RepositoryFormatKnitPack3,
2810
RepositoryFormatPackDevelopment0Subtree,
2812
return (isinstance(source._format, nosubtrees) and
2813
isinstance(target._format, subtrees))
2814
except AttributeError:
2818
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2819
"""See InterRepository.fetch()."""
2820
from bzrlib.fetch import Knit1to2Fetcher
2821
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2822
self.source, self.source._format, self.target,
2823
self.target._format)
2824
f = Knit1to2Fetcher(to_repository=self.target,
2825
from_repository=self.source,
2826
last_revision=revision_id,
2827
pb=pb, find_ghosts=find_ghosts)
2828
return f.count_copied, f.failed_revisions
2831
class InterDifferingSerializer(InterKnitRepo):
2834
def _get_repo_format_to_test(self):
2838
def is_compatible(source, target):
2839
"""Be compatible with Knit2 source and Knit3 target"""
2840
if source.supports_rich_root() != target.supports_rich_root():
2842
# Ideally, we'd support fetching if the source had no tree references
2843
# even if it supported them...
2844
if (getattr(source, '_format.supports_tree_reference', False) and
2845
not getattr(target, '_format.supports_tree_reference', False)):
2850
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2851
"""See InterRepository.fetch()."""
2852
revision_ids = self.target.search_missing_revision_ids(self.source,
2853
revision_id, find_ghosts=find_ghosts).get_keys()
2854
revision_ids = tsort.topo_sort(
2855
self.source.get_graph().get_parent_map(revision_ids))
2856
def revisions_iterator():
2857
for current_revision_id in revision_ids:
2858
revision = self.source.get_revision(current_revision_id)
2859
tree = self.source.revision_tree(current_revision_id)
2861
signature = self.source.get_signature_text(
2862
current_revision_id)
2863
except errors.NoSuchRevision:
2865
yield revision, tree, signature
2867
my_pb = ui.ui_factory.nested_progress_bar()
2872
install_revisions(self.target, revisions_iterator(),
2873
len(revision_ids), pb)
2875
if my_pb is not None:
2877
return len(revision_ids), 0
2880
class InterOtherToRemote(InterRepository):
2882
def __init__(self, source, target):
2883
InterRepository.__init__(self, source, target)
2884
self._real_inter = None
2887
def is_compatible(source, target):
2888
if isinstance(target, remote.RemoteRepository):
2892
def _ensure_real_inter(self):
2893
if self._real_inter is None:
2894
self.target._ensure_real()
2895
real_target = self.target._real_repository
2896
self._real_inter = InterRepository.get(self.source, real_target)
2898
def copy_content(self, revision_id=None):
2899
self._ensure_real_inter()
2900
self._real_inter.copy_content(revision_id=revision_id)
2902
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2903
self._ensure_real_inter()
2904
return self._real_inter.fetch(revision_id=revision_id, pb=pb,
2905
find_ghosts=find_ghosts)
2908
def _get_repo_format_to_test(self):
2912
class InterRemoteToOther(InterRepository):
2914
def __init__(self, source, target):
2915
InterRepository.__init__(self, source, target)
2916
self._real_inter = None
2919
def is_compatible(source, target):
2920
if not isinstance(source, remote.RemoteRepository):
2922
# Is source's model compatible with target's model?
2923
source._ensure_real()
2924
real_source = source._real_repository
2925
if isinstance(real_source, remote.RemoteRepository):
2926
raise NotImplementedError(
2927
"We don't support remote repos backed by remote repos yet.")
2928
return InterRepository._same_model(real_source, target)
2930
def _ensure_real_inter(self):
2931
if self._real_inter is None:
2932
self.source._ensure_real()
2933
real_source = self.source._real_repository
2934
self._real_inter = InterRepository.get(real_source, self.target)
2936
def fetch(self, revision_id=None, pb=None, find_ghosts=False):
2937
self._ensure_real_inter()
2938
return self._real_inter.fetch(revision_id=revision_id, pb=pb,
2939
find_ghosts=find_ghosts)
2941
def copy_content(self, revision_id=None):
2942
self._ensure_real_inter()
2943
self._real_inter.copy_content(revision_id=revision_id)
2946
def _get_repo_format_to_test(self):
2951
InterRepository.register_optimiser(InterDifferingSerializer)
2952
InterRepository.register_optimiser(InterSameDataRepository)
2953
InterRepository.register_optimiser(InterWeaveRepo)
2954
InterRepository.register_optimiser(InterKnitRepo)
2955
InterRepository.register_optimiser(InterModel1and2)
2956
InterRepository.register_optimiser(InterKnit1and2)
2957
InterRepository.register_optimiser(InterPackRepo)
2958
InterRepository.register_optimiser(InterOtherToRemote)
2959
InterRepository.register_optimiser(InterRemoteToOther)
2962
class CopyConverter(object):
2963
"""A repository conversion tool which just performs a copy of the content.
2965
This is slow but quite reliable.
2968
def __init__(self, target_format):
2969
"""Create a CopyConverter.
2971
:param target_format: The format the resulting repository should be.
2973
self.target_format = target_format
2975
def convert(self, repo, pb):
2976
"""Perform the conversion of to_convert, giving feedback via pb.
2978
:param to_convert: The disk object to convert.
2979
:param pb: a progress bar to use for progress information.
2984
# this is only useful with metadir layouts - separated repo content.
2985
# trigger an assertion if not such
2986
repo._format.get_format_string()
2987
self.repo_dir = repo.bzrdir
2988
self.step('Moving repository to repository.backup')
2989
self.repo_dir.transport.move('repository', 'repository.backup')
2990
backup_transport = self.repo_dir.transport.clone('repository.backup')
2991
repo._format.check_conversion_target(self.target_format)
2992
self.source_repo = repo._format.open(self.repo_dir,
2994
_override_transport=backup_transport)
2995
self.step('Creating new repository')
2996
converted = self.target_format.initialize(self.repo_dir,
2997
self.source_repo.is_shared())
2998
converted.lock_write()
3000
self.step('Copying content into repository.')
3001
self.source_repo.copy_content_into(converted)
3004
self.step('Deleting old repository content.')
3005
self.repo_dir.transport.delete_tree('repository.backup')
3006
self.pb.note('repository converted')
3008
def step(self, message):
3009
"""Update the pb by a step."""
3011
self.pb.update(message, self.count, self.total)
3023
def _unescaper(match, _map=_unescape_map):
3024
code = match.group(1)
3028
if not code.startswith('#'):
3030
return unichr(int(code[1:])).encode('utf8')
3036
def _unescape_xml(data):
3037
"""Unescape predefined XML entities in a string of data."""
3039
if _unescape_re is None:
3040
_unescape_re = re.compile('\&([^;]*);')
3041
return _unescape_re.sub(_unescaper, data)
3044
class _VersionedFileChecker(object):
3046
def __init__(self, repository):
3047
self.repository = repository
3048
self.text_index = self.repository._generate_text_key_index()
3050
def calculate_file_version_parents(self, text_key):
3051
"""Calculate the correct parents for a file version according to
3054
parent_keys = self.text_index[text_key]
3055
if parent_keys == [_mod_revision.NULL_REVISION]:
3057
return tuple(parent_keys)
3059
def check_file_version_parents(self, texts, progress_bar=None):
3060
"""Check the parents stored in a versioned file are correct.
3062
It also detects file versions that are not referenced by their
3063
corresponding revision's inventory.
3065
:returns: A tuple of (wrong_parents, dangling_file_versions).
3066
wrong_parents is a dict mapping {revision_id: (stored_parents,
3067
correct_parents)} for each revision_id where the stored parents
3068
are not correct. dangling_file_versions is a set of (file_id,
3069
revision_id) tuples for versions that are present in this versioned
3070
file, but not used by the corresponding inventory.
3073
self.file_ids = set([file_id for file_id, _ in
3074
self.text_index.iterkeys()])
3075
# text keys is now grouped by file_id
3076
n_weaves = len(self.file_ids)
3077
files_in_revisions = {}
3078
revisions_of_files = {}
3079
n_versions = len(self.text_index)
3080
progress_bar.update('loading text store', 0, n_versions)
3081
parent_map = self.repository.texts.get_parent_map(self.text_index)
3082
# On unlistable transports this could well be empty/error...
3083
text_keys = self.repository.texts.keys()
3084
unused_keys = frozenset(text_keys) - set(self.text_index)
3085
for num, key in enumerate(self.text_index.iterkeys()):
3086
if progress_bar is not None:
3087
progress_bar.update('checking text graph', num, n_versions)
3088
correct_parents = self.calculate_file_version_parents(key)
3090
knit_parents = parent_map[key]
3091
except errors.RevisionNotPresent:
3094
if correct_parents != knit_parents:
3095
wrong_parents[key] = (knit_parents, correct_parents)
3096
return wrong_parents, unused_keys
3099
def _old_get_graph(repository, revision_id):
3100
"""DO NOT USE. That is all. I'm serious."""
3101
graph = repository.get_graph()
3102
revision_graph = dict(((key, value) for key, value in
3103
graph.iter_ancestry([revision_id]) if value is not None))
3104
return _strip_NULL_ghosts(revision_graph)
3107
def _strip_NULL_ghosts(revision_graph):
3108
"""Also don't use this. more compatibility code for unmigrated clients."""
3109
# Filter ghosts, and null:
3110
if _mod_revision.NULL_REVISION in revision_graph:
3111
del revision_graph[_mod_revision.NULL_REVISION]
3112
for key, parents in revision_graph.items():
3113
revision_graph[key] = tuple(parent for parent in parents if parent
3115
return revision_graph